@ionivetech/mugiwara 0.5.5 → 0.6.1

package/evals/cases/routing-lessons.json

@@ -1,25 +0,0 @@
-{
-  "name": "routing-lessons",
-  "skill": "mugiwara-lessons",
-  "trigger": {
-    "positive": [
-      { "prompt": "capture what we learned from this mission", "top_k": 3 },
-      { "prompt": "read past lessons before starting the new mission", "top_k": 3 },
-      { "prompt": "update the cross-mission lessons ledger with new insights", "top_k": 3 }
-    ],
-    "negative": [
-      { "prompt": "log this error in the application monitoring system" },
-      { "prompt": "write a retrospective document for the team" }
-    ]
-  },
-  "behavioral": [
-    {
-      "task": "The mission is complete. Capture lessons learned and append them to the cross-mission ledger.",
-      "rubric": [
-        "reads past lessons at triage",
-        "captures new lessons at closure",
-        "appends to the append-only ledger"
-      ]
-    }
-  ]
-}

package/evals/cases/routing-orchestration.json

@@ -1,25 +0,0 @@
-{
-  "name": "routing-orchestration",
-  "skill": "mugiwara-orchestration",
-  "trigger": {
-    "positive": [
-      { "prompt": "triage this new feature request and decide the lane size", "top_k": 3 },
-      { "prompt": "classify this mission and coordinate the execution waves", "top_k": 3 },
-      { "prompt": "captain a new mission — classify, size, dispatch, check in", "top_k": 3 }
-    ],
-    "negative": [
-      { "prompt": "implement the feature described in the ticket" },
-      { "prompt": "write code for the database migration" }
-    ]
-  },
-  "behavioral": [
-    {
-      "task": "A new mission request arrived. Classify it, size the lane, and dispatch the appropriate waves.",
-      "rubric": [
-        "performs 5-way classification before any action",
-        "sizes the lane based on sensitivity and scope",
-        "coordinates waves without implementing code"
-      ]
-    }
-  ]
-}

package/evals/cases/routing-planning.json

@@ -1,26 +0,0 @@
-{
-  "name": "routing-planning",
-  "skill": "mugiwara-planning",
-  "trigger": {
-    "positive": [
-      { "prompt": "turn this spec into an execution plan with parallel waves", "top_k": 3 },
-      { "prompt": "plan the implementation approach before writing any code", "top_k": 3 },
-      { "prompt": "break down the feature into tasks and estimate the effort", "top_k": 3 }
-    ],
-    "negative": [
-      { "prompt": "just write the code for this simple bug fix" },
-      { "prompt": "brainstorm different architecture options" }
-    ]
-  },
-  "behavioral": [
-    {
-      "task": "Turn the approved feature spec into a detailed execution plan with parallel-proof waves and context scan.",
-      "rubric": [
-        "interviews before planning",
-        "performs a full context scan",
-        "produces parallel-proof waves",
-        "scales the plan to the request scope"
-      ]
-    }
-  ]
-}

package/evals/cases/routing-pr.json

@@ -1,25 +0,0 @@
-{
-  "name": "routing-pr",
-  "skill": "mugiwara-pr",
-  "trigger": {
-    "positive": [
-      { "prompt": "push this branch and prepare the pull request description", "top_k": 3 },
-      { "prompt": "write the PR summary with a verdict file", "top_k": 3 },
-      { "prompt": "prepare the branch for review, push it and generate PR material", "top_k": 3 }
-    ],
-    "negative": [
-      { "prompt": "merge this pull request into main" },
-      { "prompt": "create a release tag for this version" }
-    ]
-  },
-  "behavioral": [
-    {
-      "task": "The mission is complete. Push the branch and prepare PR material including a verdict file.",
-      "rubric": [
-        "does a plain git push",
-        "generates a verdict file with ready PR summary",
-        "never creates the PR, merges, or deploys"
-      ]
-    }
-  ]
-}

package/evals/cases/routing-quality.json

@@ -1,25 +0,0 @@
-{
-  "name": "routing-quality",
-  "skill": "mugiwara-quality",
-  "trigger": {
-    "positive": [
-      { "prompt": "run eslint and prettier on the changed files", "top_k": 3 },
-      { "prompt": "format the code and run the linter before committing", "top_k": 3 },
-      { "prompt": "run the project test suite and check for lint errors", "top_k": 3 }
-    ],
-    "negative": [
-      { "prompt": "check if the coverage threshold is met" },
-      { "prompt": "review the diff for breaking changes" }
-    ]
-  },
-  "behavioral": [
-    {
-      "task": "After the checkpoint passes, run the formatter, linter, and unit tests for the project.",
-      "rubric": [
-        "discovers the project's lint and format tooling",
-        "runs the unit test suite",
-        "never weakens or changes existing configs"
-      ]
-    }
-  ]
-}

package/evals/cases/routing-ship.json

@@ -1,26 +0,0 @@
-{
-  "name": "routing-ship",
-  "skill": "mugiwara-ship",
-  "trigger": {
-    "positive": [
-      { "prompt": "run the pre-launch checklist and give a go no-go decision", "top_k": 3 },
-      { "prompt": "prepare the staged rollout plan with a rollback strategy", "top_k": 3 },
-      { "prompt": "verify feature flags and ship checklist before deployment", "top_k": 3 }
-    ],
-    "negative": [
-      { "prompt": "push the branch to GitHub" },
-      { "prompt": "deploy the build artifact to the server" }
-    ]
-  },
-  "behavioral": [
-    {
-      "task": "The mission is complete and all gates passed. Run the pre-launch checklist and give a GO/NO-GO.",
-      "rubric": [
-        "runs the pre-launch checklist",
-        "checks feature flags and staged rollout plan",
-        "requires a mandatory rollback plan",
-        "gives a binary GO/NO-GO verdict"
-      ]
-    }
-  ]
-}

package/evals/cases/routing-sunset.json

@@ -1,25 +0,0 @@
-{
-  "name": "routing-sunset",
-  "skill": "mugiwara-sunset",
-  "trigger": {
-    "positive": [
-      { "prompt": "remove the old v1 API endpoints and retire them safely", "top_k": 3 },
-      { "prompt": "delete the legacy payment module and migrate to v2", "top_k": 3 },
-      { "prompt": "deprecate this old endpoint and phase it out over two releases", "top_k": 3 }
-    ],
-    "negative": [
-      { "prompt": "add a new version of the payment API" },
-      { "prompt": "refactor the v2 endpoint handler for performance" }
-    ]
-  },
-  "behavioral": [
-    {
-      "task": "Retire the legacy v1 payment endpoints. They have been replaced by v2 for 6 months.",
-      "rubric": [
-        "runs the keep-or-retire gate before any removal",
-        "plans a phased cutover with safe DB migrations",
-        "every removal has a tested way back"
-      ]
-    }
-  ]
-}

package/evals/cases/routing-using-mugiwara.json

@@ -1,25 +0,0 @@
-{
-  "name": "routing-using-mugiwara",
-  "skill": "using-mugiwara",
-  "trigger": {
-    "positive": [
-      { "prompt": "how do I use mugiwara to accomplish a task?", "top_k": 3 },
-      { "prompt": "what crew members are available?", "top_k": 3 },
-      { "prompt": "route this task to the right specialist", "top_k": 3 }
-    ],
-    "negative": [
-      { "prompt": "what does git status do?" },
-      { "prompt": "run npm install for me" }
-    ]
-  },
-  "behavioral": [
-    {
-      "task": "Explain how mugiwara works and route a simple task.",
-      "rubric": [
-        "explains the crew structure",
-        "classifies the task correctly",
-        "routes to the right specialist"
-      ]
-    }
-  ]
-}

package/evals/cases/routing-workflow.json

@@ -1,25 +0,0 @@
-{
-  "name": "routing-workflow",
-  "skill": "mugiwara-workflow",
-  "trigger": {
-    "positive": [
-      { "prompt": "start a new mission for this feature — run the full pipeline", "top_k": 3 },
-      { "prompt": "run the complete workflow from triage to closure", "top_k": 3 },
-      { "prompt": "we have a non-trivial task, kick off the 9-wave pipeline", "top_k": 3 }
-    ],
-    "negative": [
-      { "prompt": "fix this one-line typo in the README" },
-      { "prompt": "just run the tests and close the mission" }
-    ]
-  },
-  "behavioral": [
-    {
-      "task": "A new non-trivial feature request arrived. Run the full mugiwara pipeline from triage to closure.",
-      "rubric": [
-        "starts with Luffy triage gateway",
-        "sequences through brainstorm/plan/execute/checkpoint/quality/gates/review/heal/closure waves",
-        "does not skip waves for non-trivial missions"
-      ]
-    }
-  ]
-}

package/evals/floor.json

@@ -1,6 +0,0 @@
-{
-  "rank1": 88.2,
-  "topk": 98.7,
-  "negatives": 100,
-  "updated": "2026-08-11"
-}