{"_id":"@adnova-group/muster","_rev":"20-835ec07a4b4d76c0b47a3e601241e676","name":"@adnova-group/muster","dist-tags":{"latest":"0.7.1"},"versions":{"0.1.0":{"name":"@adnova-group/muster","version":"0.1.0","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.1.0","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://github.com/Adnova-Group/muster#readme","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"b42622419fdce84bdcfb5efbb5ec63c7b330cb4d","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.1.0.tgz","fileCount":153,"integrity":"sha512-8APPg8cthOQAX+1x0TgT4Ql7esULqdD8sIPnLnPBZoyWYgomEHtim0Jh4IgnKkh27aMShXm/sw0wNp2bU/CCOQ==","signatures":[{"sig":"MEUCIQCQyF9dT8Z5ZpWMN3XdouFIku2eGtHNbrzfuvIfHofyhgIgHVZD40MJlypr2yS25Qerph6BjIzlhjDdwKMMuQSDOtw=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":802469},"type":"module","engines":{"node":">=20"},"gitHead":"8cf99c2df6d18c5a9972105ee7c7560d43a69122","scripts":{"test":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"deprecated":"plugin prose invokes the wrong npx package; upgrade to 0.2.4","repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.1.0_1780938894560_0.23806258820904258","host":"s3://npm-registry-packages-npm-production"}},"0.2.0":{"name":"@adnova-group/muster","version":"0.2.0","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.2.0","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://github.com/Adnova-Group/muster#readme","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"ac090eb536092e2e959662cc7dc34b152910e489","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.2.0.tgz","fileCount":155,"integrity":"sha512-6T+tG82YHSbSUkzEnw3+iFMY5kbJR1iYhft/8yvvkQTnm+wO0YZvVONItfXHrEAXb/CEJvysZIHsn8gK3geDJA==","signatures":[{"sig":"MEQCIGDtXyCSjoPgh13XV1QRkTaXrFcup47TLLw+0qc94AJiAiBxF0QPu8rU2rZYL/yE4RFzYkEffBlJ9FrEXC9EK0tv3g==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":810863},"type":"module","engines":{"node":">=20"},"gitHead":"2cce3083104fa17812ac88fa4af30a31e01dfd29","scripts":{"test":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"deprecated":"plugin prose invokes the wrong npx package; upgrade to 0.2.4","repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.2.0_1780943953435_0.8220613452141039","host":"s3://npm-registry-packages-npm-production"}},"0.2.1":{"name":"@adnova-group/muster","version":"0.2.1","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.2.1","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"37dcb2cce38bb0510a49613fa9e965a1e046c256","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.2.1.tgz","fileCount":156,"integrity":"sha512-hO2SbHSMA+rVaGXk8xUHJnAfDmGpLMFSD1wp30Y7fpXcohX1kO64ye1br0QPjowTX3DZTqN2fR0kdaWTNKzmjw==","signatures":[{"sig":"MEQCIAZkb/ubvikT2dkPrOtsFTXuC2gyQt48GsULES/StdhNAiA6gsavojJLeAg4VdHxdy+ZPokxbFA84oMLcCUoGpTkFQ==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":816386},"type":"module","engines":{"node":">=20"},"gitHead":"9e7969ae146df03d00a4c0fa3eee9a45d989cba7","scripts":{"test":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"deprecated":"plugin prose invokes the wrong npx package; upgrade to 0.2.4","repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.2.1_1780947518612_0.5127136411075552","host":"s3://npm-registry-packages-npm-production"}},"0.2.2":{"name":"@adnova-group/muster","version":"0.2.2","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.2.2","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"02d72168ff1a3edf9c645e46c92a1e2bf1e9a587","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.2.2.tgz","fileCount":156,"integrity":"sha512-5+Pm2yTT2KXTUWFmYEtOi7GqIaM5743pRId8AoI98Rs78vvUcIyWRCw0X0+5Jufpj1Cygq2fA+aNmnKMSTGCAA==","signatures":[{"sig":"MEUCIFn+RpeYF7AGhSyulH4B3wdkHy9uQqZAkfCLsds0eKJpAiEAwUiN4HZ+pK8mPNjlgx6HZdE+N7GEJ/sjq2Gteeh+bhE=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":819967},"type":"module","engines":{"node":">=20"},"gitHead":"2b6cf32f952e810e15bdd8dc47d480d38291612b","scripts":{"test":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"deprecated":"plugin prose invokes the wrong npx package; upgrade to 0.2.4","repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.2.2_1780954317242_0.7808661292164936","host":"s3://npm-registry-packages-npm-production"}},"0.2.3":{"name":"@adnova-group/muster","version":"0.2.3","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.2.3","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"a8ff432b1b2a87aedd5f8e654007c7adf10ac89c","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.2.3.tgz","fileCount":158,"integrity":"sha512-WjXzxzPz3NeSDbyF26g873bEToNdKxoi94pX201NIk9HLseVAtvPBgjPuW1c7vwFhOBnfe5l/lu3c43EuTDxwg==","signatures":[{"sig":"MEQCIGL/t1duoKdRgPY+L7273HkdoG5V9MzkVg7xpfg9hD3/AiAJR5HGpyVEFozdl+dYBHid12rXsg87QiAOtWgnR324aw==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":826227},"type":"module","engines":{"node":">=20"},"gitHead":"4bc77500967ede8ecc869877b99704206c394f0e","scripts":{"test":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"deprecated":"plugin prose invokes the wrong npx package; upgrade to 0.2.4","repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.2.3_1781026692386_0.44731496285078043","host":"s3://npm-registry-packages-npm-production"}},"0.2.4":{"name":"@adnova-group/muster","version":"0.2.4","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.2.4","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"7adb39ec8d1b115ef6c592e59e331ee85bf81c62","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.2.4.tgz","fileCount":161,"integrity":"sha512-pI9HsY7+wmVHFOXk9E+VN7xGV7EJVBNq92uMQcwoBqWvgMr3qmzq9bJXiOwK3imq6JJ1D1CQ3xM1BT2RVzGUKQ==","signatures":[{"sig":"MEYCIQD4gFra9JhjKGE8AvQuoMnFyEuIqVeKT3iZPHw1sDtlJgIhAKfIKtZfQmO9MACAMhhe6Iwjs+LSmY5x+53W169R6+q9","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":866414},"type":"module","engines":{"node":">=20"},"gitHead":"73373727a4fb3417df5ccb24e09b681c651451f6","scripts":{"test":"node --test","prepublishOnly":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.2.4_1781059103443_0.7759101096441174","host":"s3://npm-registry-packages-npm-production"}},"0.2.5":{"name":"@adnova-group/muster","version":"0.2.5","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.2.5","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"ea9c3b9031e677e78032876a23c09df34ceb0364","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.2.5.tgz","fileCount":161,"integrity":"sha512-XC8lx+x0EjLunEFThxBmlSI8YCKWNMAptgKd4s+AzmSTmmKZaibizwGYJT8WIOVucrKk1rzU3rjqb0RMoAJczg==","signatures":[{"sig":"MEQCIB8opEEcDsDS3+06lhEEVS41t6gvIMiOGtJ/FoXpdY6ZAiABY+DcKlHXBJH6VKuHbelAeFxnl64teWiwb1h76w84+A==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":869062},"type":"module","engines":{"node":">=20"},"gitHead":"a4a9cdc6933d828a0b89a1f2c5ef495c55384ed7","scripts":{"test":"node --test","prepublishOnly":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.2.5_1781124946241_0.47177956767492524","host":"s3://npm-registry-packages-npm-production"}},"0.2.7":{"name":"@adnova-group/muster","version":"0.2.7","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.2.7","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"54c2444ead9253a1bf487ea37e57f046914ea7a5","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.2.7.tgz","fileCount":164,"integrity":"sha512-awCD6X4Rs14tgQIoj30hPDgG88aelT0K65BipdafNX7WSTwVIKNt8ziN6w/vli0Y0JXe39ZOwfrJpsE72cC+CQ==","signatures":[{"sig":"MEUCIFDwfDAjvGyR5sKQHTtoc4jEPS6ULBq3dh1mijmaWKLTAiEAjKYe2POCrKHAlsau0yaypP8EBj6t0sPD1FSltK/YlmY=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":905937},"type":"module","engines":{"node":">=20"},"gitHead":"84861b292cae96991a7b1af22b349255c2e023d1","scripts":{"test":"node --test","prepublishOnly":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.2.7_1781556763706_0.23285541752923367","host":"s3://npm-registry-packages-npm-production"}},"0.2.8":{"name":"@adnova-group/muster","version":"0.2.8","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.2.8","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"5259b61d35d315d9881534a0c0e971dafd017ee7","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.2.8.tgz","fileCount":169,"integrity":"sha512-hviveMRwzd7ogohORTuqqbRdbjrLJPU9X9HjvVrOMxUBd22YziWUFZKtH49V1jxgmFFi8CNdnFygmfQ1xcz/zQ==","signatures":[{"sig":"MEUCIQCaw8b2XapUmtSfZtbXNbrPvj0PFTKTCqgKmrxPBQ4JOgIgM6aebJxK8VIfD1THPGd5A7MvIUMK+NAFzaBe2r8Wb/8=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":981370},"type":"module","engines":{"node":">=20"},"gitHead":"d6205749814dd2a26eefcac8e26353676e953619","scripts":{"test":"node --test","eval:router":"node eval/router/grade.mjs","prepublishOnly":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.2.8_1781813520984_0.8169318266524743","host":"s3://npm-registry-packages-npm-production"}},"0.3.0":{"name":"@adnova-group/muster","version":"0.3.0","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.3.0","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"a285df261fa43fdd7dcecfd58debfde7c8870008","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.3.0.tgz","fileCount":171,"integrity":"sha512-rAPN+rY5OxBOg9iY9kfyxQSlHWVQRsaYWlwg+lP97m8yKf82+54IRPo7i9Q3+Wz4/9PePlBwnL4TIHs1FHPjug==","signatures":[{"sig":"MEQCIC5QNeY8juwvQRA3kus/SFU564F9+t4oVErTNTWGu/o0AiAQ0clFCZ80J5vQnRSfuVc7iDQ4yeOIwIl2Ao36BrwLbw==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":1016757},"type":"module","engines":{"node":">=20"},"gitHead":"a8d7d970c7bee9f1e685528730fcdbdd868b87a1","scripts":{"test":"node --test","eval:router":"node eval/router/grade.mjs","prepublishOnly":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.3.0_1782245434711_0.8792738255082564","host":"s3://npm-registry-packages-npm-production"}},"0.3.1":{"name":"@adnova-group/muster","version":"0.3.1","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.3.1","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"9576b247c3c9927985e7b11aef167583ba72d67f","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.3.1.tgz","fileCount":175,"integrity":"sha512-GuQX/0vMZxKBvb22+SL5DdLYOZi7gnkBso8O7Nhi/bpNBKpD8FwQLdjHWlBa1o4rjku2urDUatjVll98YPpeEg==","signatures":[{"sig":"MEQCIENRzYx6SzqpdLpX4cFtWIP2M7EHALuFITBnsbIPr85xAiBPFz35fSO24S+NSsuh79ISmuiYDlPagIdlKNxAdg2yJg==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":1049907},"type":"module","engines":{"node":">=20"},"gitHead":"46258b30d4a7a41f45888ace1943d0c6181696a0","scripts":{"test":"node --test","eval:router":"node eval/router/grade.mjs","prepublishOnly":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.3.1_1782949147113_0.7712590827401993","host":"s3://npm-registry-packages-npm-production"}},"0.3.2":{"name":"@adnova-group/muster","version":"0.3.2","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.3.2","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"697985189ad1263314d1a8355ae36545f0d12196","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.3.2.tgz","fileCount":179,"integrity":"sha512-98r0OllBU10gz7vwcjV7Pg7bNWzrhi/XvdhaxyfIGxoQFLxsZuAOcBKCwawlxZIE2x2Lx+5Rz24W60VDlgDEDA==","signatures":[{"sig":"MEYCIQDxG66i5qUuFGqmKV+rPBwW4WMx26hTjPuF4NoYhQG86gIhAMySwhvdOVMTaQ7ecA8+yCdDW2+i/Gm7Ua0FngCYtUvY","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":1082041},"type":"module","engines":{"node":">=20"},"gitHead":"ab93bda8a6595f1b6bf7ac07f93a1807570a5170","scripts":{"test":"node --test","eval:router":"node eval/router/grade.mjs","prepublishOnly":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.3.2_1783009891094_0.6352765894792807","host":"s3://npm-registry-packages-npm-production"}},"0.3.3":{"name":"@adnova-group/muster","version":"0.3.3","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.3.3","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"b15b40baa1684b3ff39027f2c8a72e4428b87df0","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.3.3.tgz","fileCount":180,"integrity":"sha512-au40e95GJODfcTxRd2XYD7iLq2ZrORpk2l/gNaMTJR7hEsPaNCa3+BTL/nvRrKb9l06JB9WnmDFjWZizi57yQg==","signatures":[{"sig":"MEQCICldpMlUm0rGDozQU5OzmJ05BLt2ngZTfkok/WcyWRMVAiBJwOXdiVYVPfiWvXZTd4op2KVFlDsaWOUBlpdDHc1Mqw==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":1096585},"type":"module","engines":{"node":">=20"},"gitHead":"db65a0fa443eb1e7393b9be879a9269f125e840e","scripts":{"test":"node --test","eval:router":"node eval/router/grade.mjs","prepublishOnly":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.3.3_1783199923950_0.7712065805182242","host":"s3://npm-registry-packages-npm-production"}},"0.4.0":{"name":"@adnova-group/muster","version":"0.4.0","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.4.0","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"1eeec663d02b7a8c10c0d7f40a35481fc5f2626a","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.4.0.tgz","fileCount":192,"integrity":"sha512-6igBccmy0D6vCWbwxKjNk1ZbBjhrOeRbjc1vFQo3Ghu15cCd3/LnCiIGN8tPyFKqAOBtWKLx+Qrzp0Di8FplKw==","signatures":[{"sig":"MEUCICImdkSA9fIHDfCfNAe/5I/EBfOIrQKwQ3i7JtTroFo1AiEA3mvv048pDDqDvenMXp9gLKY9qfb0xkmUZJMjltzJy04=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":1300359},"type":"module","engines":{"node":">=20"},"gitHead":"bbf6ad7d8cce09ec2b62b9f19407f08e09e3c25e","scripts":{"test":"node --test","eval:modes":"node eval/modes/grade.mjs","eval:router":"node eval/router/grade.mjs","prepublishOnly":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.4.0_1783385332436_0.49887446918552625","host":"s3://npm-registry-packages-npm-production"}},"0.4.1":{"name":"@adnova-group/muster","version":"0.4.1","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.4.1","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"cbd96ffc397d74d0b0949ee1de9891b3c5617955","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.4.1.tgz","fileCount":200,"integrity":"sha512-3UquRcK9IUd0YSt34JEqww+Kt0sBdzbSC+n8OcpsCfDqK6goYjCHA/RR0OQcwdGjWhbV3ENJoLlMW5Qk+CxRTw==","signatures":[{"sig":"MEUCIHAI9VeVGnBZ1v7yGbITZWT4LDvRdfXfG2BLYi39LmDnAiEAzG4s4q7lzGiAufVwKTIS/IvY+CCik0LNOdGf8y8Fw5s=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":1406538},"type":"module","engines":{"node":">=20"},"gitHead":"c151ef35ff2f7a7cd4c273b0037f1373323c9e17","scripts":{"test":"node --test","eval:modes":"node eval/modes/grade.mjs","eval:router":"node eval/router/grade.mjs","prepublishOnly":"node --test"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.4.1_1783472939405_0.20426415558698574","host":"s3://npm-registry-packages-npm-production"}},"0.5.0":{"name":"@adnova-group/muster","version":"0.5.0","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.5.0","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"57efae4692937dcd1fa2a1d13d19acdc40b903d6","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.5.0.tgz","fileCount":311,"integrity":"sha512-4r5wZA1bnGxQ3UfNeqQAU8gq/kFsydukkZQrPaubos4OS7Fo8pweSlXf4pF5FG2o0Zshvfq4qkYOK13TO3vrTA==","signatures":[{"sig":"MEUCIQC0Kt4U7XkDHwhq/YQx6rwdSlMaRplnHT7EEmCo59ItoAIgS/VCZJ2CjaoVESsTqGZQGjVoQY6JvOKEUoM/Tc0MoZY=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":2753505},"type":"module","engines":{"node":">=20"},"gitHead":"f49152e029c6f4badc6ca8d45710b55af5d45ee8","scripts":{"test":"node --test","pretest":"node scripts/build-codex.mjs","eval:modes":"node eval/modes/grade.mjs","build:codex":"node scripts/build-codex.mjs","check:codex":"node scripts/check-codex.mjs","eval:router":"node eval/router/grade.mjs","prepublishOnly":"npm run build:codex && npm run check:codex && node --test","sync:codex-assets":"node scripts/sync-codex-skill-assets.mjs"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0","esbuild":"^0.25.12"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.5.0_1784595454969_0.9177456504919281","host":"s3://npm-registry-packages-npm-production"}},"0.6.0":{"name":"@adnova-group/muster","version":"0.6.0","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.6.0","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"7701ed75eebb7cffdf358428ad683944e2ad37d3","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.6.0.tgz","fileCount":361,"integrity":"sha512-j+cuay5ujcl2wuP64YOCduGh9DLw8MUJox/bDFUbVeuHQvUGHLCO+V9DMwJvIKi2e24M6C6HlEi7nhDUGQ38yg==","signatures":[{"sig":"MEUCIQDrBolATEs/CQ3RGAHJE2OyGpOfCMF8GEA3SBo8N1sqjQIgcepSZnOv2Om8VM4Nmk9Mf+4acNV3oXdsRbVeINzaQBI=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":3514191},"type":"module","engines":{"node":">=20"},"gitHead":"edddfa31821af974e5a7cbd3709d855b5dbf8a18","scripts":{"test":"node --test","pretest":"node scripts/build-codex.mjs","eval:modes":"node eval/modes/grade.mjs","build:codex":"node scripts/build-codex.mjs","check:codex":"node scripts/check-codex.mjs","eval:router":"node eval/router/grade.mjs","prepublishOnly":"npm run build:codex && npm run check:codex && node --test","sync:codex-assets":"node scripts/sync-codex-skill-assets.mjs"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0","esbuild":"^0.25.12"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.6.0_1785541794234_0.7413279716939658","host":"s3://npm-registry-packages-npm-production"}},"0.7.0":{"name":"@adnova-group/muster","version":"0.7.0","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"author":{"name":"Adnova Group"},"license":"Apache-2.0","_id":"@adnova-group/muster@0.7.0","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"homepage":"https://adnova-group.github.io/muster/","bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"bin":{"muster":"src/cli.js"},"dist":{"shasum":"3a4b17829756302f020f72f5a626fd35bd635754","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.7.0.tgz","fileCount":417,"integrity":"sha512-03wybwaxUbUCN5hZNGgLPZnpoHGteyGLUZrG/0GM5WzLf0GhviAkN8UMHOyhZLJFRkt72Pu6KQhEPwUauj6eFA==","signatures":[{"sig":"MEYCIQCZ83T5CP701XBJ+ARvuowh2uV5tc7bxPEhtavKLVdhGAIhANVgsUqBD4nh3A6i48Y2ZWnNk6WetIyGiLhh8n8Mu/NF","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":4317270},"type":"module","engines":{"node":">=20"},"gitHead":"12a2b6eca0f0762f44a0ebd57ef06f4bccf39000","scripts":{"test":"node scripts/test.mjs test/*.test.js","pretest":"node scripts/build-codex.mjs","eval:modes":"node eval/modes/grade.mjs","build:codex":"node scripts/build-codex.mjs","check:codex":"node scripts/check-codex.mjs","eval:router":"node eval/router/grade.mjs","prepublishOnly":"npm run build:codex && npm run check:codex && node --test","sync:codex-assets":"node scripts/sync-codex-skill-assets.mjs"},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"repository":{"url":"git+https://github.com/Adnova-Group/muster.git","type":"git"},"_npmVersion":"11.12.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","directories":{},"_nodeVersion":"26.0.0","dependencies":{"yaml":"^2.5.0","esbuild":"^0.25.12"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/muster_0.7.0_1786394517708_0.13748470721770678","host":"s3://npm-registry-packages-npm-production"}},"0.7.1":{"name":"@adnova-group/muster","version":"0.7.1","description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","license":"Apache-2.0","author":{"name":"Adnova Group"},"type":"module","publishConfig":{"access":"public"},"homepage":"https://adnova-group.github.io/muster/","repository":{"type":"git","url":"git+https://github.com/Adnova-Group/muster.git"},"bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"bin":{"muster":"src/cli.js"},"engines":{"node":">=20"},"scripts":{"test":"node scripts/test.mjs test/*.test.js","pretest":"node scripts/build-codex.mjs","eval:router":"node eval/router/grade.mjs","eval:modes":"node eval/modes/grade.mjs","build:codex":"node scripts/build-codex.mjs","check:codex":"node scripts/check-codex.mjs","sync:codex-assets":"node scripts/sync-codex-skill-assets.mjs","prepublishOnly":"npm run build:codex && npm run check:codex && node --test"},"dependencies":{"esbuild":"^0.25.12","yaml":"^2.5.0"},"gitHead":"9273ffd375fc59e2882906d2048e722cce4dcdad","_id":"@adnova-group/muster@0.7.1","_nodeVersion":"26.0.0","_npmVersion":"11.12.1","dist":{"integrity":"sha512-Mmwp3ZKh5zS5MafLamGlux09XsFBn5uf7FP2eVAuKWfWhk6Z62NijZ6vJ5RLkS59Hh8EVDMVQFFy36EfVCj76A==","shasum":"d20f950ed873345ce9c53eeae1ac080df66161a6","tarball":"https://registry.npmjs.org/@adnova-group/muster/-/muster-0.7.1.tgz","fileCount":417,"unpackedSize":4332266,"signatures":[{"keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U","sig":"MEYCIQCTBRR9LAc3Id7qVK/TF4atUAy1xSV3pH0tNcxs5866jQIhAPsgoEUH4CC5UlugcEC+LPfhPWVN517REcvjUnCF+ltw"}]},"_npmUser":{"name":"adnova-group","email":"ryan@adnovagroup.com"},"directories":{},"maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"_npmOperationalInternal":{"host":"s3://npm-registry-packages-npm-production","tmp":"tmp/muster_0.7.1_1786581351343_0.05882618663765693"},"_hasShrinkwrap":false}},"time":{"created":"2026-06-08T17:14:54.359Z","modified":"2026-08-13T00:35:51.702Z","0.1.0":"2026-06-08T17:14:54.728Z","0.2.0":"2026-06-08T18:39:13.634Z","0.2.1":"2026-06-08T19:38:38.816Z","0.2.2":"2026-06-08T21:31:57.410Z","0.2.3":"2026-06-09T17:38:12.574Z","0.2.4":"2026-06-10T02:38:23.613Z","0.2.5":"2026-06-10T20:55:46.396Z","0.2.7":"2026-06-15T20:52:43.899Z","0.2.8":"2026-06-18T20:12:01.160Z","0.3.0":"2026-06-23T20:10:34.876Z","0.3.1":"2026-07-01T23:39:07.327Z","0.3.2":"2026-07-02T16:31:31.450Z","0.3.3":"2026-07-04T21:18:44.186Z","0.4.0":"2026-07-07T00:48:52.649Z","0.4.1":"2026-07-08T01:08:59.667Z","0.5.0":"2026-07-21T00:57:35.170Z","0.6.0":"2026-07-31T23:49:54.444Z","0.7.0":"2026-08-10T20:41:57.917Z","0.7.1":"2026-08-13T00:35:51.537Z"},"bugs":{"url":"https://github.com/Adnova-Group/muster/issues"},"author":{"name":"Adnova Group"},"license":"Apache-2.0","homepage":"https://adnova-group.github.io/muster/","keywords":["orchestrator","agentic","ai","claude-code","cli","multi-agent","pipelines","glass-box"],"repository":{"type":"git","url":"git+https://github.com/Adnova-Group/muster.git"},"description":"Glass-box agentic orchestrator: detects context, assembles the right crew, and shows its reasoning.","maintainers":[{"name":"adnova-group","email":"ryan@adnovagroup.com"}],"readme":"# Muster\n\nGlass-box, multi-domain agentic orchestrator for Claude Code, Codex, Kimi, and Cowork. Give it an outcome; it assembles the right crew and shows its reasoning before it acts.\n\n[![License: Apache-2.0](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](LICENSE)\n[![Node >= 20](https://img.shields.io/badge/node-%3E%3D20-brightgreen.svg)](package.json)\n[![npm](https://img.shields.io/npm/v/@adnova-group/muster.svg)](https://www.npmjs.com/package/@adnova-group/muster)\n[![Docs](https://img.shields.io/badge/docs-adnova--group.github.io%2Fmuster-6d5ce7.svg)](https://adnova-group.github.io/muster/)\n\n**Documentation: [adnova-group.github.io/muster](https://adnova-group.github.io/muster/)**\n\n## What it is\n\nMuster turns an outcome into finished work. It detects your project, discovers the capabilities you already have installed, picks the best tool for each piece of the job, and runs a crew of specialists toward your success criteria. Every decision is inspectable: which role resolved to which provider, on which model, and why.\n\nIt runs on Claude Code, Codex, Kimi, or Cowork with no separate model API, and it gets better as you install more tools. Runtime support differs: Claude Code and Codex expose the full native command surface, Kimi uses namespaced native skills, and Cowork's verified MCP lane has a seven-mode protocol subset. ChatGPT Work is a separate, conditional local/private MCP lane: configuration alone is not proof that the active Work host invoked Muster. The work is not limited to code. Product, business, content, and operations are first-class.\n\n## Quickstart\n\n```sh\nnpx -y @adnova-group/muster@0.7.1 install\n```\n\nPin the reviewed release in automation and copy the current version from `package.json` when updating. `npx` uses npm's execution path and may download the named package from the configured registry before running it. Review the package provenance and release notes before changing the pin.\n\n`install` mutates nothing in your `~/.claude`. It only prints the steps it cannot do for you, because registering a plugin is a Claude Code action:\n\n```sh\n/plugin marketplace add Adnova-Group/muster  # register the marketplace\n/plugin install muster@muster                # install the plugin\n```\n\nMuster's glass-box output style ships inside the plugin and applies automatically when the plugin is enabled (`force-for-plugin`), so there is no command to run. Plugin install is a Claude Code action, so the running session picks Muster up only after you (re)install it through `/plugin`. Restart or `/clear`, then run your first outcome:\n\n```\n/muster:plan Add rate limiting to the public API with tests\n```\n\n### Codex CLI and Desktop\n\nBuild or install the package, then install Muster's managed Codex profiles and plugin:\n\n```sh\nnpx -y @adnova-group/muster@0.7.1 install codex --scope project\n```\n\n`--scope project` writes Muster-owned profiles and declarations under the project's `.codex/` layer, installs the hook runtime under `.codex/muster/`, and merges owned hook groups into `.codex/hooks.json`. The install also records ownership receipts and registers the plugin. Existing unrelated profiles, configuration, and hook groups are preserved.\n\nCodex CLI, Desktop, and the IDE share `$CODEX_HOME/config.toml`. Even for `--scope project`, the installer ensures the canonical `agents.max_concurrent_threads_per_session` setting exists, defaulting it to `12` only when the user has not configured a canonical or legacy ceiling. Existing positive user ceilings are preserved. A receipt lets the last managed-scope uninstall restore only values Muster changed, including cleanup of legacy `max_threads`/`max_depth` values only when an older Muster receipt proves ownership. The user scope is canonical for hooks: a healthy user install makes a project install skip its own hook merge, avoiding duplicate events. Use `--dry-run` to inspect the complete write, merge, registration, and cleanup plan first:\n\n```sh\nnpx -y @adnova-group/muster@0.7.1 install codex --scope project --dry-run\n```\n\nOn Windows, install from the same host that runs Codex. Native Windows and WSL have different home directories, Node installations, plugin caches, and normally different `CODEX_HOME` values; an install under `~/.codex` in WSL does not configure native Codex Desktop under `%USERPROFILE%\\.codex`. If you switch hosts, rerun the scoped install and `muster doctor --codex` in the new host. Do not point native Desktop at a WSL-only path or assume that a `/mnt/c/...` checkout makes the two user scopes identical.\n\nCodex requires a new trust review when installed hook definitions change. Inspect exact definitions with `/hooks`; update-sensitive trust means a previously trusted hash does not authorize changed code. To remove only Muster-owned project state, preview and then run:\n\n```sh\nnpx -y @adnova-group/muster@0.7.1 uninstall codex --scope project --dry-run\nnpx -y @adnova-group/muster@0.7.1 uninstall codex --scope project\n```\n\nUninstall preserves unrelated config and Codex's project trust records, removes only receipted Muster declarations and hook groups, prunes Muster-owned hook trust entries, and unregisters the plugin only after the last managed scope is gone. With Codex on `PATH`, install registers `Adnova-Group/muster` and adds `muster@muster` idempotently. Without Codex it installs profiles and hooks, then prints the exact registration follow-up.\n\nUse `$muster` or a mode skill such as `$muster-plan`, `$muster-go`, `$muster-audit`, or `$muster-capture`. Codex users can inspect live Codex capability state with `muster capabilities --codex` and run `muster doctor --codex`.\n\nFrom an interactive terminal, `muster codex-plan \"<outcome>\"` starts a new local App Server session, discovers the advertised Plan preset, invokes the installed `muster-plan` skill with `turn/start.collaborationMode`, and reports the effective mode from `thread/settings/updated`. It relays Muster's structured approval question to that terminal, but never overrides the user's approval policy, reviewer, permissions, or sandbox. It cannot switch an already-running desktop/IDE chat. Non-interactive callers and unavailable or unconfirmed App Server control exit nonzero with the safe in-session path: `/plan $muster-plan <outcome>`.\n\nThe Codex plugin bundles the deterministic CLI, all pipelines, 31 MCP tools, 27 custom-agent profiles, and 84 skills: 11 public mode/router skills plus 73 internal skills (22 native orchestration skills and 51 capability skills). The npm installer adds Codex-native lifecycle hooks through the supported project or user `hooks.json` layer, and the Codex plugin itself is deliberately hooks-free so the two never double-fire. These hooks fail open and are diagnostic: they cannot reliably block every unified-shell or subagent action. Todo and spawn enforcement remain advisory, and write-capable waves must use isolated Git worktrees.\n\n### Kimi\n\nKimi installs the same ten primary modes as namespaced `/muster-*` skills. In-session native subagents are the supported execution path. `kimi-process-dispatch` remains descriptor-only, and the attended `kimi-process-run` lane is report-only on every platform until Muster can bootstrap a trusted immutable, kernel-bound broker. Filesystem dispatch receipts and worktree paths are diagnostic only; they never authorize signaling a process.\n\n### ChatGPT Work (private/local plugin lane)\n\nChatGPT Work is a private/local, proof-gated plugin lane: it does not inherit Codex Desktop's `AGENTS.md`, skills, hooks, MCP, or `config.toml`. Muster's MCP surface has one neutral core in `mcp/server.mjs` with an explicit Work host adapter; `cowork/chatgpt-work-server.mjs` remains a compatibility shim, and the public Work runtime command stays `runtime/chatgpt-work-server.mjs`. Installing the connection, running Secure MCP Tunnel, choosing a profile (`pro-safe` or `full`), and the native-proof contract are covered in the [ChatGPT Work guide](https://adnova-group.github.io/muster/guides/chatgpt-work).\n\n## The ten modes\n\n| Mode | Command | What it does |\n| --- | --- | --- |\n| Plan | `/muster:plan <outcome \\| backlog text>` | Approve-first entry point: confirms outcome-vs-backlog scope, shows the glass-box crew manifest, then stops for Approve & run / Adjust / Cancel. A confirmed backlog scope hands off to Plan-backlog. |\n| Go | `/muster:go <outcome \\| backlog text>` | Hands-off entry point: the same scope confirm as Plan, then runs the full lifecycle unattended (branch, route, waves, merge) and stops only for the merge decision or an escalation. |\n| Plan-backlog | `/muster:plan-backlog <backlog ref \\| raw intent>` | Approve-first batch planner: routes every backlog item up front into one batch plan and stops for approval; Approve & clear hands off to Go-backlog. |\n| Go-backlog | `/muster:go-backlog <backlog ref>` | The batch clearer: runs every ready item to disposition, sequentially or in dependency waves when annotated, with one attended stop at the end. |\n| Diagnose | `/muster:diagnose <symptom>` | Failure-first bug fix: reproduce, find root cause, fix, add a regression test, verify. No symptom-patching. |\n| Audit | `/muster:audit [path]` | Breadth-first whole-codebase review and fix across six dimensions (seven when the project builds prompts or agents), then fixes everything with tests and verifies. |\n| Design | `/muster:design <action>` | Resolve canonical `DESIGN.md` context and digest receipts, initialize it with an attended hold, inspect bounded design evidence/provider state, or run one of 23 pinned design workflows. |\n| Runner | `/muster:runner [source]` | Unattended one-cycle work-picker for a Claude Code Routine or cron: resumes an answered blocked item or claims exactly one available item, drives it through the full Go lifecycle force-coerced to a `pr` disposition, leaves a receipt, and stops. The schedule provides the loop, not the verb. |\n| Capture | `/muster:capture [hint]` | Conversation-to-backlog generator: mines the session's discussion (findings, decisions, review residuals, an explicit directive) into backlog items via the same extract/validate/dedupe/write machinery, gated by your approval before anything is written. Writes only `.muster/backlog.md` -- it never assembles a crew or runs work itself. |\n| Init | `/muster:init [dir]` | Prepare a repository profile and receipt, hand native instruction work to the active runtime, and finalize only after positive evidence or an acknowledged unavailable handoff. |\n\nThe mechanics behind this table, including the approval-gate details, task fences (`owns`/`frozen`/`forbiddenActions`), the manifest's `mergeDisposition`, and the wave-mode schedule, are documented in the [mode reference](https://adnova-group.github.io/muster/reference/modes).\n\nDesign mode's context resolution, attended initialization, bounded detector, provider contract, and complete workflow list are documented in the [Design reference](https://github.com/Adnova-Group/muster/blob/main/docs/design.md).\n\nPlan and Go accept a GitHub issue reference (a bare number, `#123`, or an issues URL) as the outcome; both also accept the same backlog refs as Plan-backlog and Go-backlog (a backlog `.md` path, `issues:<label>`, or `linear:<key>`) and confirm the scope before planning a whole batch. A thin outcome gets refined first: `muster assess` does a deterministic gap-check, and if the outcome is vague, an interview skill asks one question at a time behind an approval gate before any crew is assembled. An outcome that decomposes into independent parts can instead be written to a backlog (`.muster/backlog.md`) for `/muster:go-backlog` to clear as a batch, or for `/muster:plan-backlog` to batch-plan first; `/muster:audit backlog [path]` fills the same backlog from audit's findings, sweeping read-only instead of fixing them inline; `/muster:capture [hint]` fills it a third way, mining a conversation's findings and decisions instead of an audit sweep or an interview decomposition.\n\nAn item annotated with `{id}`/`{deps}` switches `/muster:go-backlog` into wave mode. Every ready item is eligible for an isolated build and review regardless of whether its eventual disposition is `pr`, `keep`, `merge-local`, or `merge-push`. Concurrency is bounded by the declared environment and wave size (`MUSTER_SPRINT_PARALLEL`, default 5 and hard ceiling 10), not by a fixed three-runner rule. After every worker wake, the orchestrator drains all available completion receipts, reconciles once, dispatches every newly eligible action, and reconciles again before it may wait. The wave-wide build/review barrier must pass before ordered disposition/integration begins, and the next dependency wave waits for that ordered lane. A harness without safe parallel dispatch executes the same schedule sequentially without changing dependency or integration order. Go-backlog and Runner share a **coordination** skill (claim/receipt/ledger discipline) so a scheduled Runner and an attended Go-backlog clear can safely work the same backlog or `issues:<label>` at once.\n\n## Initialize a repository\n\nUse Init before the greenfield workflow or when adopting Muster in a cloned repository:\n\n```sh\n/muster:init [dir]\n# or\nnpx -y @adnova-group/muster init [dir]\n```\n\nInit performs bounded, read-only project learning and writes only the owned pair `.muster/project-profile.json` and `.muster/init-receipt.json`. Both are schema-versioned canonical JSON files. The profile records classification, repository facts, and a SHA-256 state fingerprint. It is provider/model-neutral: it never stores provider IDs, concrete model names, resolved roles, capability inventories, or timestamps. Same-state reruns return the same receipt without rewriting bytes. A pending bare rerun only observes expected native artifacts and does not change the baseline or state.\n\nFor an empty greenfield directory, preparation may safely initialize `.git` before creating the owned pair. This is the one trust-boundary mutation: Git uses a fresh empty template and controlled built-ins, with no repository hooks or discovered commands. Native instruction generation stays with the active runtime. Claude Code and Codex handoffs use one canonical instruction pair: `AGENTS.md` is authoritative, and `CLAUDE.md` contains exactly:\n\n```md\n# Claude Code\n\n@AGENTS.md\n```\n\nIf conflicting instruction files existed at the preparation baseline, Init does not overwrite or merge them; it leaves a HUMAN-HOLD for the user to reconcile. Init records a `not-requested` state, then the runtime-specific workflow moves it to `handoff` or a proven callable adapter moves it to `attempted`. A native `/init` request, suggestion, command invocation, refusal to overwrite, or existing artifact alone is not completion. Completion requires artifact-delta evidence, an explicit confirmation of a pre-existing artifact, or a bounded call-result receipt. Do not pass `--evidence-file` with `artifact-delta`; that flag is only for `preexisting-confirmed` and `call-result`. If the handoff is unavailable, including Copilot or an unknown runtime without a proven adapter, Init leaves a HUMAN-HOLD; run `muster init acknowledge [dir] --reason unavailable` to permit finalization while the native state remains `handoff`. Never invent a native command for an unavailable runtime.\n\nGreenfield finalization may create only a missing `.gitignore`, `README.md`, and `.gitkeep` files under the `docs/design` and `docs/plan` directories. Brownfield finalization creates none of those seeds and preserves the clone's README, docs, instruction files, hooks, settings, and other user content. Init never executes repository setup instructions, package scripts, hooks, dependency installers, or discovered commands. The explicit legacy `muster setup [dir]` command still scaffolds its older seed set; new greenfield guidance uses Init and keeps the design-before-plan-before-implementation gate.\n\n## How it works\n\nThe novel core is a capability and domain router. Muster names a fixed vocabulary of roles (the kinds of work a crew might need), and each role resolves through a ladder, best available first:\n\n1. An installed external provider (a plugin, agent, or MCP server you already have)\n2. A Muster built-in agent\n3. A Muster built-in skill\n4. Inline (the model does it directly)\n\n`muster capabilities` walks this ladder for every role and reports the winner, the full fallback chain, installable recommendations, and the chosen model. Because the chain always ends at inline, every role resolves to something, so Muster works on bare Claude Code and improves as you add tools.\n\nThe role set is fixed but the provider set is not. When an outcome does not fit a named role, description-search bridges the gap: `muster match \"<task>\"` ranks every catalog provider by deterministic token overlap (no model call), so \"audit this code for security vulnerabilities\" surfaces the security specialist even though it never names a role.\n\nEach role also carries a conceptual tier picked to fit the work. The ladder, from least to most capable, is `scout`, `core`, `prime`, and `apex`. Mechanical roles use scout, routine work defaults to core, and peak-judgment roles use apex with a deterministic fallback to prime. Runtime adapters map those tiers to concrete models. Muster composes the tools you already have and falls back to its own. For the full design, see the [architecture reference](https://adnova-group.github.io/muster/reference/architecture) (or the [architecture source](https://github.com/Adnova-Group/muster/blob/main/docs/architecture.md)).\n\n## Claude Code-only lifecycle hooks\n\nThis section describes the plugin-native Claude Code hooks. Codex uses the diagnostic hook layer described above. Other harnesses receive only the lifecycle primitives they actively subscribe to; Muster does not infer an active harness subscription from files on disk.\n\nClaude Code receives four plugin-native hooks. Enforcement follows the run's external effects, not the orchestrator's own in-repo edits: the only hard deny on a tool call is the action-class fence, scoped to a live run that declared a forbidden action. The fence classifies every payload before considering delegation or metadata-path exemptions, so a recognized forbidden external effect remains denied when a call carries `agent_id` or points at `.muster/`/`.claude/`. It fails open when run markers are absent or unreadable, classification is ambiguous, or no forbidden class matches. Delegated and metadata-path calls are exempt only from the advisory inline-drift accounting after they clear the fence; dispatch briefs, worktree barriers, and review gates remain required. A second, narrower block lives on `TaskCompleted`, gating the native task board's own completion tick rather than a tool call. Everything else is a warn-only border invitation.\n\n- **`SessionStart`**: injects a one-line pointer (muster available; `/muster:plan` for orchestration-scale work) at the start of every session. On a fresh start it clears per-session drift state, but preserves repository-global `.muster/run-active`/`wave-active` markers whose activity lease is fresh or cannot be verified; only a regular marker idle for more than one hour is recovered as stale. `PreToolUse` renews the lease during active work.\n- **`UserPromptSubmit`**: the only prompt-time nudge is the isDirective-triggered border invitation. A directive-shaped prompt is eligible only after inline-file activity corroborates orchestration scale; a cold \"fix typo\" does not trigger it. One invitation starts a shared 15-minute cooldown. The signal re-arms after a Muster run starts, `SessionStart`, or 60 minutes of inactivity, but the cooldown still prevents rapid repeats.\n- **`PreToolUse`**: the action-class fence (the one hard deny) plus the tool-call half of the same border invitation. While a muster run is active AND `.muster/forbidden-actions` lists a class, a tool call classified into that class (send/sign/submit/publish/purchase/delete-remote) is denied, honoring `MUSTER_ACTION_GUARD` (`off`/`warn`/deny-by-default). Independently, a cumulative counter of distinct inline-edited files (across turns, with no muster run active) crossing `MUSTER_INLINE_SCALE` (default 3) warns once per crossing with the same value-toned copy -- never denies. Writes into `.muster/` and `.claude/` (in-cwd repo) are exempt from this advisory drift accounting only after action classification.\n- **`TaskCompleted`**: the second block surface. The orchestrator writes `.muster/task-board.json` (one entry per native task muster created) and flips an entry to `reviewGate: \"pass\"` only once the review gate actually passes that task. This hook denies (exit 2) a completion tick on a tracked task that has no recorded PASS, so a task cannot be marked done before it has been reviewed; it fails open for anything the board does not track. `MUSTER_TASK_GATE=off` disables it.\n\nAll four hooks live inside the plugin, so they activate when muster is enabled and go away when muster is disabled. They do not write to your `~/.claude/CLAUDE.md` or `settings.json` and create no global files. Each hook is fail-safe: any error falls back to an empty result and never blocks a session from starting.\n\n## Pipelines\n\nA pipeline is a phased, gated recipe for producing one kind of artifact. Each declares a domain, an ordered list of phases, and a gate. Gating uses a floor principle: the weakest dimension must clear the floor and the total must clear a pass threshold, so a strong average cannot rescue one weak dimension.\n\nThe set spans software and knowledge work. A few examples: PRD, business-case, launch-plan, executive-summary, OKRs, AI implementation spec, competitive-battlecard, blog-post, case-study, runbook, video-content, and book (fiction and non-fiction). Roadmap prioritization is one to call out: goals go in, and a RICE-ranked now/next/later roadmap comes out, with the model estimating the factors and the CLI doing the arithmetic. Human-facing pipelines end with a humanize phase that strips em-dashes, AI-tell words, and robotic cadence. Content pipelines that name an audience or a voice resolve a named profile from `docs/profiles/AUDIENCES.md`/`VOICE.md` (creating or extending it on first use) and calibrate depth/jargon/altitude and register/rhythm to it; a `docs/profiles/BRAND.md` anchors image-prompt and publish-phase visuals to a shared palette.\n\n## Prompt evaluation\n\nMuster can lint, eval, and optimize prompts, including the prompts an application generates at runtime to build agents and agentic workflows, and prompts found in a codebase Muster is working in. The deterministic core runs offline; a built-in skill (`muster-prompt-smith`, the `prompt-quality` role) supplies the model calls for empirical eval.\n\n- **Lint** (`muster prompt lint <file|->`) is a no-LLM structural check that enforces Anthropic's best practices (role, XML tags, multishot examples, explicit output format, positive framing) and the agent/guardrail rules (imperative tool framing, stop conditions, \"I don't know\" allowance, citations, input separation). Every finding cites the doc rule it comes from and suggests a fix; the rubric is gated by the same floor principle as pipelines. Pass `--agent --tools` for runtime agent prompts.\n- **Eval** (`muster prompt eval <suite.json>`) grades outputs against a test dataset with code graders (`json`/`regex`/`python`) plus an LLM-judge, and reports accuracy.\n- **Optimize** (`muster prompt variations` then `muster prompt optimize`) generates technique-driven variations, re-scores them, and keeps the winner via the tournament floor, flagging a regression when nothing beats the baseline.\n- **Scan** (`muster prompt scan <dir>`) walks a repo for prompts (markdown skill/agent/command docs, `.prompt` files, code assignments) and lints each; it powers the conditional `prompt-quality` audit dimension.\n\nThe linter is genre-aware (`--system` relaxes task-only rules for instruction prompts), ignores code fences across languages, and lets a prompt opt out of a rule inline (`<!-- prompt-lint-disable RULE: reason -->`). A prompt with zero findings scores a perfect 15/15.\n\n```sh\n# lint a runtime agent prompt piped straight from your app\nyour-app --print-agent-prompt | npx -y @adnova-group/muster prompt lint - --agent --tools\n```\n\nSee the [commands reference](https://adnova-group.github.io/muster/reference/commands) for the full surface.\n\n## Configuration\n\nMuster's runtime behavior can be tuned with environment variables:\n\n| Variable | Default | Semantics |\n| --- | --- | --- |\n| `MUSTER_INLINE_SCALE` | `3` | The border-invitation threshold: the Nth distinct file edited inline across turns, with no muster run active, crosses the border and warns once per crossing (never denies). |\n| `MUSTER_MAX_TIER` | _(unset)_ | Caps the conceptual model tier policy. For example, `MUSTER_MAX_TIER=prime` excludes apex and `MUSTER_MAX_TIER=core` enables budget mode; unset means no cap. Static agent frontmatter pins are not affected on direct invocation; in Muster runs the dispatch override honors the cap. |\n| `MUSTER_ENABLE_APEX` | _(unset)_ | Enables apex for peak-judgment roles such as the tournament judge, architecture review, improve, and advisor. Unset (or `0`/`false`) degrades apex to prime deterministically; `1`/`true` enables it when the harness can serve it. |\n| `MUSTER_ACTION_GUARD` | `deny` | Action-class fence on `PreToolUse` while `.muster/forbidden-actions` is present: `deny` blocks a matching send/sign/submit/publish/purchase/delete-remote tool call, `warn` allows with a reminder, `off` disables the fence. This is the only hard-deny surface left in muster's enforcement stack. |\n| `MUSTER_ADVISOR_MAX_CONSULTS` | `3` | Maximum advisor consults per run. Bounds the cost of workers escalating to the advisor role. Set to 0 to disable advisor consults. |\n| `MUSTER_FUSE_TOPK` | `3` | Maximum number of tournament candidates passed to the fusion synthesizer. Must be >= 1. |\n| `MUSTER_FUSE_MIN_DISAGREEMENT` | `1` | Minimum disagreement score required to activate fusion synthesis. Below this threshold `muster fuse` falls back to the single best candidate. Set to 0 to always fuse when >= 2 candidates pass. |\n| `MUSTER_SPRINT_PARALLEL` | `5` | Max concurrent item-runner subagents per wave in `/muster:go-backlog` wave mode; hard ceiling `10` (higher values clamp, `0` is invalid; concurrency is never unbounded). Read by go-backlog's orchestration protocol, not by library code. |\n\nThis table covers the most common controls. See the [full configuration reference](https://adnova-group.github.io/muster/reference/configuration) for hook cooldown, model, concurrency, and advanced runtime settings.\n\n<!-- legacy-tier-compat:start -->\nCompatibility only: legacy input aliases `haiku`, `sonnet`, `opus`, and `fable` normalize to `scout`, `core`, `prime`, and `apex`; they are not a second conceptual ladder. `MUSTER_ENABLE_FABLE` remains an alias for `MUSTER_ENABLE_APEX`.\n<!-- legacy-tier-compat:end -->\n\n## Built on\n\nMuster's design was inspired by atomic-claude, superpowers, and gsd-core. It vendors a curated set of MIT-licensed skills and agents, with every source and item recorded for attribution:\n\n| Source | License | Provides |\n| --- | --- | --- |\n| obra/superpowers | MIT | Brainstorming, planning, TDD, code-review, debugging, verification skills |\n| wshobson/agents | MIT | Software and knowledge-work agents across many specialties |\n| open-gsd/gsd-core | MIT | Plan, execute, and verify workflow phases |\n\nFor Codex, Muster prefers enabled authoritative upstream implementations when they exist: the official Superpowers plugin, WSHObson's per-plugin Codex skills, and GSD's installer-generated Codex skills. Bundled `sp-*`, `wsh-*`, and `muster-gsd-*` skills remain deterministic fallbacks and never install those providers implicitly. The pinned compatibility survey, including Atomic Codex, Book Genesis, humanizer sources, and Promptfoo, is recorded in [`codex/upstreams.json`](codex/upstreams.json).\n\nAlongside the vendored material, Muster ships its own clean-room specialists, authored fresh from the role concept. Full provenance lives in [NOTICE](NOTICE).\n\n## Contributing and license\n\nContributions are welcome. See the [contribution guide](https://github.com/Adnova-Group/muster/blob/main/CONTRIBUTING.md) to get started.\n\nMuster is licensed under Apache-2.0. See [LICENSE](LICENSE) and [NOTICE](NOTICE).\n","readmeFilename":"README.md"}