@ionivetech/mugiwara 0.5.0 → 0.5.2

package/evals/cases/routing-contract-first.json

@@ -0,0 +1,25 @@
+{
+  "name": "routing-contract-first",
+  "skill": "mugiwara-contract-first",
+  "trigger": {
+    "positive": [
+      { "prompt": "design the API contract for the new payment service", "top_k": 3 },
+      { "prompt": "define the interface and error semantics before implementing", "top_k": 3 },
+      { "prompt": "make sure this API change is backward compatible", "top_k": 3 }
+    ],
+    "negative": [
+      { "prompt": "write the implementation for the payment handler" },
+      { "prompt": "add a database migration for the new field" }
+    ]
+  },
+  "behavioral": [
+    {
+      "task": "Design the API contract for a new service before any implementation starts.",
+      "rubric": [
+        "defines the contract before any implementation",
+        "specifies error semantics and boundary validation",
+        "checks backward compatibility with existing consumers"
+      ]
+    }
+  ]
+}

package/evals/cases/routing-execution.json

@@ -0,0 +1,25 @@
+{
+  "name": "routing-execution",
+  "skill": "mugiwara-execution",
+  "trigger": {
+    "positive": [
+      { "prompt": "execute the approved plan step by step", "top_k": 3 },
+      { "prompt": "implement the todo list from the plan", "top_k": 3 },
+      { "prompt": "carry out the tasks in parallel batches and commit each one", "top_k": 3 }
+    ],
+    "negative": [
+      { "prompt": "let's brainstorm the architecture first" },
+      { "prompt": "review the plan for gaps before we start" }
+    ]
+  },
+  "behavioral": [
+    {
+      "task": "Execute the approved plan: 3 parallel tasks covering auth middleware, database migration, and API handlers.",
+      "rubric": [
+        "processes the todo list sequentially",
+        "runs independent tasks in parallel batches",
+        "commits per logical task with evidence"
+      ]
+    }
+  ]
+}

package/evals/cases/routing-frontend.json

@@ -0,0 +1,26 @@
+{
+  "name": "routing-frontend",
+  "skill": "mugiwara-frontend",
+  "trigger": {
+    "positive": [
+      { "prompt": "redesign the dashboard page with responsive layout", "top_k": 3 },
+      { "prompt": "make this form accessible with proper ARIA labels", "top_k": 3 },
+      { "prompt": "create a reusable button component from the design tokens", "top_k": 3 }
+    ],
+    "negative": [
+      { "prompt": "add a REST endpoint for the user dashboard data" },
+      { "prompt": "design the database schema for user settings" }
+    ]
+  },
+  "behavioral": [
+    {
+      "task": "Redesign the settings page component with responsive breakpoints and WCAG 2.1 AA compliance.",
+      "rubric": [
+        "matches existing repo component patterns",
+        "extracts design tokens from the existing system",
+        "implements responsive behavior",
+        "meets WCAG 2.1 AA accessibility standards"
+      ]
+    }
+  ]
+}

package/evals/cases/routing-gates.json

@@ -0,0 +1,25 @@
+{
+  "name": "routing-gates",
+  "skill": "mugiwara-gates",
+  "trigger": {
+    "positive": [
+      { "prompt": "check if the coverage threshold is met before we merge", "top_k": 3 },
+      { "prompt": "verify the definition of done — build, lint, tests all pass", "top_k": 3 },
+      { "prompt": "verify coverage thresholds, build exit code, and definition of done", "top_k": 3 }
+    ],
+    "negative": [
+      { "prompt": "run the formatter and linter on the changed files" },
+      { "prompt": "review the diff for stylistic issues" }
+    ]
+  },
+  "behavioral": [
+    {
+      "task": "The quality checks finished. Now verify if the coverage thresholds are met and the build passes.",
+      "rubric": [
+        "checks coverage thresholds with binary verdicts",
+        "verifies build exit code is zero",
+        "does not negotiate or weaken thresholds"
+      ]
+    }
+  ]
+}

package/evals/cases/routing-git.json

@@ -0,0 +1,25 @@
+{
+  "name": "routing-git",
+  "skill": "mugiwara-git",
+  "trigger": {
+    "positive": [
+      { "prompt": "commit these changes with atomic commits per logical unit", "top_k": 3 },
+      { "prompt": "split this diff into separate commits by topic", "top_k": 3 },
+      { "prompt": "find which commit introduced this bug using bisect", "top_k": 3 }
+    ],
+    "negative": [
+      { "prompt": "push the branch and open a pull request" },
+      { "prompt": "deploy the release to production" }
+    ]
+  },
+  "behavioral": [
+    {
+      "task": "Split the staged changes into atomic commits: one for the refactor, one for the new feature, one for the tests.",
+      "rubric": [
+        "creates atomic commits per logical change",
+        "detects the repo's commit style",
+        "does not push or open a PR"
+      ]
+    }
+  ]
+}

package/evals/cases/routing-healing.json

@@ -0,0 +1,25 @@
+{
+  "name": "routing-healing",
+  "skill": "mugiwara-healing",
+  "trigger": {
+    "positive": [
+      { "prompt": "the previous execution wave failed, fix what is broken", "top_k": 3 },
+      { "prompt": "there are blockers in the ledger, triage and fix them", "top_k": 3 },
+      { "prompt": "the blocker ledger has entries, stop the line and fix them", "top_k": 3 }
+    ],
+    "negative": [
+      { "prompt": "execute the next wave in the plan" },
+      { "prompt": "add a new feature to the backlog" }
+    ]
+  },
+  "behavioral": [
+    {
+      "task": "The test suite failed after the last wave. The blocker ledger has 3 entries. Fix the failures.",
+      "rubric": [
+        "reads the blocker ledger before taking action",
+        "applies stop-the-line triage",
+        "proves the fix before moving to next wave"
+      ]
+    }
+  ]
+}

package/evals/cases/routing-lessons.json

@@ -0,0 +1,25 @@
+{
+  "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

@@ -0,0 +1,25 @@
+{
+  "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

@@ -0,0 +1,26 @@
+{
+  "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

@@ -0,0 +1,25 @@
+{
+  "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-proof-order.json

@@ -0,0 +1,25 @@
+{
+  "name": "routing-proof-order",
+  "skill": "mugiwara-proof-order",
+  "trigger": {
+    "positive": [
+      { "prompt": "write the test first before implementing the function", "top_k": 3 },
+      { "prompt": "use TDD — red green refactor for this feature", "top_k": 3 },
+      { "prompt": "prove the behavior with a failing test, then implement", "top_k": 3 }
+    ],
+    "negative": [
+      { "prompt": "run the existing unit test suite that was written last week" },
+      { "prompt": "run the acceptance tests defined by the user" }
+    ]
+  },
+  "behavioral": [
+    {
+      "task": "Implement a new utility function. Write the test first, then the implementation, then refactor.",
+      "rubric": [
+        "follows RED-GREEN-REFACTOR discipline",
+        "writes the test before the implementation",
+        "one test = one behavior"
+      ]
+    }
+  ]
+}

package/evals/cases/routing-quality.json

@@ -0,0 +1,25 @@
+{
+  "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

@@ -0,0 +1,26 @@
+{
+  "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

@@ -0,0 +1,25 @@
+{
+  "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-workflow.json

@@ -0,0 +1,25 @@
+{
+  "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

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

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ionivetech/mugiwara",
-  "version": "0.5.0",
+  "version": "0.5.2",
   "description": "The Straw Hat crew of AI agents and skills: brainstorm, plan, execute, checkpoint, quality, gates, review, security, self-healing. Installs into Claude Code, opencode, Copilot, Gemini, Codex, Cursor, Kimi, pi, Windsurf, Cline, Kilo, Antigravity.",
   "type": "module",
   "license": "MIT",
@@ -43,6 +43,7 @@
     "evals": "bun scripts/run-evals.ts",
     "retrieval-eval": "bun scripts/retrieval-eval.ts",
     "sync-version": "bun scripts/sync-version.ts",
+    "gate": "bun run typecheck && bun run test && bun run build && bun scripts/validate-content.ts --check-manifest --check-docs && bun scripts/run-evals.ts && bun scripts/retrieval-eval.ts",
     "prepack": "bun run build && bun run sync-version"
   },
   "keywords": [

package/scripts/probe.ts

@@ -0,0 +1,40 @@
+#!/usr/bin/env bun
+// Ad-hoc retrieval probe. Usage: bun scripts/probe.ts "your prompt here"
+
+import { readdirSync, readFileSync, existsSync, statSync } from 'node:fs';
+import { join } from 'node:path';
+import { parseFrontmatter } from '../src/frontmatter.ts';
+
+const skillsDir = join(import.meta.dirname, '..', 'content', 'skills');
+const STOP = ['the', 'and', 'for', 'use', 'when', 'that', 'with', 'this', 'from', 'its', 'not', 'are', 'has'];
+const tok = (t: string) => t.toLowerCase().replace(/[^a-z0-9\s-]/g, ' ')
+  .split(/\s+/).filter(x => x.length > 1 && !STOP.includes(x));
+
+const terms = new Map<string, number>();
+const docs = new Map<string, Map<string, number>>();
+for (const d of readdirSync(skillsDir).filter(d => statSync(join(skillsDir, d)).isDirectory())) {
+  const f = join(skillsDir, d, 'SKILL.md');
+  if (!existsSync(f)) continue;
+  const { data } = parseFrontmatter(readFileSync(f, 'utf8'));
+  const tf = new Map<string, number>();
+  for (const t of tok(data.description ?? '')) {
+    tf.set(t, (tf.get(t) || 0) + 1);
+    terms.set(t, (terms.get(t) || 0) + 1);
+  }
+  docs.set(d, tf);
+}
+
+const N = docs.size;
+const prompt = process.argv.slice(2).join(' ');
+const ranked = [...docs.keys()].map(d => {
+  let s = 0;
+  for (const t of tok(prompt)) {
+    const tf = docs.get(d)!.get(t) || 0;
+    if (tf) s += tf * Math.log(N / (terms.get(t) || 1));
+  }
+  return { d, s };
+}).sort((a, b) => b.s - a.s);
+
+console.log(`"${prompt}"\n`);
+ranked.slice(0, 8).forEach((r, i) =>
+  console.log(`  ${String(i + 1).padStart(2)}. ${r.d.padEnd(28)} ${r.s.toFixed(2)}`));