vigiles 5.1.0 → 6.0.0

package/README.md

@@ -5,11 +5,11 @@
 <h1 align="center">vigiles</h1>
 
 <p align="center">
-  <strong>Lint &amp; test the harness your AI agent runs on.</strong>
+  <strong>Lint, test &amp; measure the harness your AI agent runs on.</strong>
 </p>
 
 <p align="center">
-  Your CLAUDE.md, hooks, and skills steer the agent — but nothing checks they're <em>true</em>, and nothing tests they <em>work</em>. vigiles does both.
+  Your CLAUDE.md, hooks, and skills steer the agent — but nothing checks they're <em>true</em>, nothing tests they <em>work</em>, and nothing measures whether they actually <em>help</em>. vigiles does all three.
 </p>
 
 <p align="center">
@@ -20,14 +20,17 @@
 
 ---
 
-`Agent = Model + Harness`. You'd never ship an app without a linter and a test
-suite — yet the harness steering your agent runs on vibes. vigiles[^name] is the
-deterministic layer for it, and does two independent things — adopt either, or both:
+`Agent = Model + Harness`. You'd never ship an app without a linter, a test suite,
+and a benchmark — yet the harness steering your agent runs on vibes. vigiles[^name]
+is the deterministic layer for it — and the only one that can **A/B-test what
+actually moves the needle on your Claude subscription**, not metered API. Three
+things, adopt any:
 
-|             |                                                                                                                                                                                                                                                  |
-| ----------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
-| **🔎 Lint** | Every file path, script, code symbol, and linter rule your CLAUDE.md cites is checked against reality — so a renamed file or a disabled rule can't silently mislead the agent. **[→](docs/verifying-instruction-files.md)**                      |
-| **🧪 Test** | Hooks, skills, and subagents are code. vigiles tests they _do their job_ — and almost all of it is **deterministic, no API key**; the real-model evals run on your **Claude subscription**, not metered tokens. **[→](docs/harness-testing.md)** |
+|                |                                                                                                                                                                                                                                                       |
+| -------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| **🔎 Lint**    | Every file path, script, code symbol, and linter rule your CLAUDE.md cites is checked against reality — so a renamed file or a disabled rule can't silently mislead the agent. **[→](docs/verifying-instruction-files.md)**                           |
+| **🧪 Test**    | Hooks, skills, and subagents are code. vigiles tests they _do their job_ — and almost all of it is **deterministic, no API key**; the real-model evals run on your **Claude subscription**, not metered tokens. **[→](docs/harness-testing.md)**      |
+| **📊 Measure** | Does that skill or plugin actually help — or just add cost? **A/B it on real tasks** and read the bill + correctness, not the vendor's headline. The same engine ranks the hyped ecosystem: **what works vs hype**. **[→](docs/measuring-skills.md)** |
 
 Pick the one that hurts today. **Works with Claude Code and Codex**
 ([`vigiles/codex`](docs/harnesses.md)), and you can
@@ -92,7 +95,16 @@ CLAUDE.md:
 File paths, scripts, and code symbols — plus linter rules across **7 catalogs**
 (the rule exists **and is enabled**). Start with one inline comment, no new files;
 step up to a typed `.spec.ts` (compiled to CLAUDE.md, compiler-grade) when you want
-it. **[Full guide →](docs/verifying-instruction-files.md)**
+it — and that same spec is what makes your skills & subagents deterministically
+**testable** (a `result()` contract → an assert, no LLM judge — see ② below), not
+just lint-able. **[Full guide →](docs/verifying-instruction-files.md)**
+
+> **Markdown is prose; a typed spec is a _program_.** Opt in and an agent that
+> leaks, exceeds its effect floor, or hands off mismatched data is a **type
+> error** — your **multi-agent pipeline won't compile if the handoffs don't line
+> up**. Others lint prose after the fact; vigiles is a **compiler for harnesses**.
+> Graduated like TypeScript's `strict` — markdown still works, type-safety is
+> opt-in. **[The moat →](research/typed-spec-moat.md)**
 
 **Same cross-reference, any plugin.** `npx vigiles scan` checks a plugin's
 _contracts_ — every subagent tool, `mcp__server__tool`, `mcp_tool` hook, hook
@@ -121,25 +133,54 @@ It goes well past _"did it fire?"_:
 
 - **Hooks block** what they must — `runHook`, or the real agent CLI via `runHarnessTest`.
 - **Skills trigger** on the right prompts and stay quiet on the wrong ones — recall _and_ precision (`measureTriggerRate`).
-- **Behaviour is good** — score a skill's output directly, or A/B it on-vs-off for the real lift over no-skill (`measure` / `runEval`, with significance testing).
+- **Behaviour is good** — score a skill's output directly, or A/B it on-vs-off for the real lift (`measure` / `runEval`, with significance testing).
 - **Safety holds** — the agent _didn't_ push to the wrong branch or hit a paid API; `interceptTools` catches the attempt so the side effect never happens.
 
-**The eval you can actually afford.** Almost every tier runs with **no model and
-no API key** — milliseconds, on every commit. The rest drive your own `claude` CLI:
+Almost every tier runs with **no model and no API key** — milliseconds, on every
+commit. Only the real-model evals need a model, and they run on your own `claude`
+CLI. **[How it works →](docs/harness-testing.md)**
+
+## ③ Measure — does it actually help, or just cost more?
+
+A skill claims "65% fewer tokens." A plugin promises "3× faster." Stars and
+vibes — **zero measurement**. vigiles A/Bs the claim on real coding tasks, the
+harness loaded exactly as it ships, and reports the **metric triple**:
+
+```typescript
+import { measureArms } from "vigiles/testing";
+
+const r = await measureArms({
+  fixture: { "in.txt": "Implement a slug helper." },
+  task: "Read in.txt, write slugify() to slug.js, explain. Stop.",
+  arms: { baseline: {}, skill: { files: { "SKILL.md": THE_SKILL } } },
+  measure: (ctx) => ({ cost: ctx.usage.costUsd, correct: check(ctx) }),
+});
+// → is the bill lower? did correctness survive? (not just "fewer output tokens")
+```
+
+- **The bill (`costUsd`)** — the honest number; it weights cache ~0.1× and output 1×, so a "saved tokens" headline can't hide behind cheap cache.
+- **The optimization target** — whatever the skill claims to move (output tokens, latency, tool calls). Verified on its own terms.
+- **The blast radius** — correctness, a deterministic 1/0. A token win that breaks the code is **not a win**.
+
+The kicker: every run is **your own `claude` CLI on your Pro/Max subscription** —
+so you can measure on every change, and the same engine powers the **ecosystem
+benchmark** ("we A/B-tested the most-hyped skills — what works vs hype"). That's
+the moat: continuous, sub-affordable measurement nobody metered-by-the-token can
+match. **[Measure a skill →](docs/measuring-skills.md)** · **[Why it's affordable →](docs/eval-architecture.md)**
+
+**The eval you can actually afford.** Most of vigiles needs no model at all; the
+rest drives your subscription:
 
 |                        | Runs on                 | Cost                                        |
 | ---------------------- | ----------------------- | ------------------------------------------- |
 | promptfoo, DeepEval, … | metered API SDK         | billed **per token, every run**             |
 | **vigiles**            | your Claude Pro/Max sub | **$0 extra** — and most tiers need no model |
 
-That's why you can eval your harness on every change, not just once.
-**[How it works →](docs/harness-testing.md)** · **[Why it's affordable →](docs/eval-architecture.md)** · **[Safety model →](docs/safety.md)**
-
 ## More
 
 - **[Plugin health leaderboard →](docs/cli.md#scan-dir)** — point `scan` at a marketplace (e.g. `wshobson/agents`) and it ranks every plugin by structural health (0–100, A–F), worst issues first — still **no key**. Add `--trigger` for the model-gated column: do the skills actually fire?
-- **[CLI & GitHub Action →](docs/cli.md)** — every command, the Action (inputs / output / versioning), the Claude Code plugin, and the `lint` rules.
-- **[Skills →](docs/skills.md)** — consumer skills installed as a Claude Code plugin: `/plugin marketplace add zernie/vigiles` then `/plugin install vigiles@vigiles` (or let `vigiles init` do it). The model-invocable ones (`test-harness`, `strengthen`, `edit-spec`) fire on their own — ask _"test my skills"_, _"strengthen my rules"_, or _"add a rule to CLAUDE.md"_ and the agent reaches for them; `migrate-to-spec` and `linter-docs` are user-invoked.
+- **[CLI & GitHub Action →](docs/cli.md)** — every command, the Action (inputs / output / versioning), and the Claude Code plugin. The full **[lint rules matrix →](docs/verifying-instruction-files.md#the-validation-rules--the-full-matrix)** lives with the linting guide.
+- **[Skills →](docs/skills.md)** — consumer skills installed as a Claude Code plugin: `/plugin marketplace add zernie/vigiles` then `/plugin install vigiles@vigiles` (or let `vigiles init` do it). The model-invocable ones (`test-harness`, `strengthen`, `edit-spec`) fire on their own — ask _"test my skills"_, _"strengthen my rules"_, or _"add a rule to CLAUDE.md"_ and the agent reaches for them; `adopt-spec` and `linter-docs` are user-invoked.
 - **[Docs index →](docs/README.md)** · **[Research →](research/README.md)** · **[Related tools →](docs/related-tools.md)** (ast-grep, Dependency Cruiser, Ruler, rulesync).
 - Companion to [Feedback Loop Is All You Need](https://zernie.com/blog/feedback-loop-is-all-you-need).
 

package/dist/adapters/claude-code/adapter.js

@@ -23,6 +23,7 @@ exports.claudeCodeAdapter = {
         referenceVerification: true,
         harnessTesting: true,
         shellHooks: true,
+        subagents: true,
     },
     dialect: dialect_js_1.claudeCodeDialect,
     layout: layout_js_1.claudeCodeLayout,

package/dist/adapters/claude-code/agent-runtime.d.ts

@@ -22,6 +22,7 @@
  * (mirrors `.vigiles/active-skill.json`). The decision logic below is
  * harness-agnostic and fully testable.
  */
+import type { PurityLevel } from "../../core/effects.js";
 /**
  * Parse an agent's allowed-tools contract from its compiled markdown.
  *
@@ -41,6 +42,14 @@ export declare function parseAgentTools(markdown: string): string[] | null;
  * rail (`tools:`) and the `disallowed-tools-contract` scan/lint.
  */
 export declare function parseAgentToolList(markdown: string, key: string): string[] | null;
+/**
+ * Parse the declared purity floor from a compiled agent's `.md` — the
+ * `<!-- vigiles:purity:LEVEL -->` marker `compile` emits (see `purityMarker`).
+ * Returns null when no marker is present (the unit declared no floor, so the
+ * purity gate imposes no constraint). The single source of truth the runtime
+ * gate reads, exactly like `tools:` for the tool-contract rail.
+ */
+export declare function parseAgentPurity(markdown: string): PurityLevel | null;
 export interface PreToolDecision {
     /** Whether the tool call is allowed (true) or blocked (false). */
     readonly allow: boolean;
@@ -64,11 +73,41 @@ export declare function clearActiveAgent(cwd: string): void;
 /** The path of the active agent's compiled `.md`, or null when none is active. */
 export declare function readActiveAgent(cwd: string): string | null;
 /**
- * PreToolUse-hook decision. If an agent is active, parse its compiled `.md`
- * tool contract and allow the call only when the tool is in the allowlist;
- * otherwise block and tell the model which tools it may use. With no active
- * agent (or an agent that inherits all tools), always allow — the rail only
- * constrains agents that declared a contract.
+ * EXPERIMENTAL — parked (P3), flat-only, do NOT auto-wire. This deterministic
+ * subagent-window tracking assumes FLAT dispatch (one active subagent at a time);
+ * it is NOT nesting-safe — Claude Code v2.1.172 added depth-5 nested subagents, so
+ * correct tracking needs a depth-aware STACK (push on dispatch, pop on SubagentStop)
+ * + verifying the spawn tool name (`Agent` vs `Task`). See
+ * research/effect-boundary-design.md ("Why dropped").
+ *
+ * Resolve a `Task` tool's `subagent_type` to the compiled agent `.md` to
+ * activate, or null when none is found. The DETERMINISTIC open signal that
+ * replaces the model-invoked `agent-start`: Claude Code fires `PreToolUse` for
+ * the parent's `Task` dispatch (and `SubagentStop` when it returns), so the
+ * harness — not the model — brackets the subagent's active window. The name is
+ * the last ":"-segment (a `--plugin-dir` subagent_type is namespaced
+ * "plugin:name"); searched in `agents/` under the cwd then the plugin root. A
+ * path under the cwd is returned relative (readActiveAgent resolves vs cwd); a
+ * plugin-root hit is absolute. Returns null on an unknown agent (fail-open: an
+ * unresolved subagent is simply not gated, exactly as before agent-start ran).
+ */
+export declare function resolveDispatchedAgent(subagentType: string, cwd: string, pluginRoot?: string): string | null;
+/**
+ * The agent `.md` to activate for a `PreToolUse(Task)` event, or null. Pure
+ * (reads `tool_input.subagent_type`, resolves via {@link resolveDispatchedAgent}).
+ */
+export declare function decideTaskDispatch(toolInput: unknown, cwd: string, pluginRoot?: string): string | null;
+/**
+ * PreToolUse-hook decision. If an agent is active, enforce BOTH deterministic
+ * rails its compiled `.md` declares, in order:
+ *
+ *  1. the tool-contract rail (`tools:`) — allow only listed tools;
+ *  2. the purity gate (`vigiles:purity:`) — allow only calls within the declared
+ *     effect floor, refining `Bash` by the live `command` (`decidePurityGate`).
+ *
+ * The first to deny wins, feeding its reason back to the model. With no active
+ * agent (or one that declared neither contract), always allow — the rails only
+ * constrain agents that opted in.
  */
-export declare function evaluatePreToolUse(cwd: string, tool: string): PreToolDecision;
+export declare function evaluatePreToolUse(cwd: string, tool: string, command?: string): PreToolDecision;
 //# sourceMappingURL=agent-runtime.d.ts.map

package/dist/adapters/claude-code/agent-runtime.js

@@ -26,14 +26,20 @@
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.parseAgentTools = parseAgentTools;
 exports.parseAgentToolList = parseAgentToolList;
+exports.parseAgentPurity = parseAgentPurity;
 exports.decidePreToolUse = decidePreToolUse;
 exports.setActiveAgent = setActiveAgent;
 exports.clearActiveAgent = clearActiveAgent;
 exports.readActiveAgent = readActiveAgent;
+exports.resolveDispatchedAgent = resolveDispatchedAgent;
+exports.decideTaskDispatch = decideTaskDispatch;
 exports.evaluatePreToolUse = evaluatePreToolUse;
 const node_fs_1 = require("node:fs");
 const node_path_1 = require("node:path");
 const frontmatter_read_js_1 = require("../../core/frontmatter-read.js");
+const effects_js_1 = require("../../core/effects.js");
+const dialect_js_1 = require("./dialect.js");
+const effect_region_js_1 = require("./effect-region.js");
 // ---------------------------------------------------------------------------
 // Parse the tool contract from a compiled agent .md
 // ---------------------------------------------------------------------------
@@ -60,6 +66,18 @@ function parseAgentTools(markdown) {
 function parseAgentToolList(markdown, key) {
     return (0, frontmatter_read_js_1.frontmatterList)((0, frontmatter_read_js_1.readFrontmatter)(markdown), key);
 }
+const PURITY_RE = /<!--\s*vigiles:purity:(pure|bounded|unrestricted)\s*-->/;
+/**
+ * Parse the declared purity floor from a compiled agent's `.md` — the
+ * `<!-- vigiles:purity:LEVEL -->` marker `compile` emits (see `purityMarker`).
+ * Returns null when no marker is present (the unit declared no floor, so the
+ * purity gate imposes no constraint). The single source of truth the runtime
+ * gate reads, exactly like `tools:` for the tool-contract rail.
+ */
+function parseAgentPurity(markdown) {
+    const m = PURITY_RE.exec(markdown);
+    return m ? m[1] : null;
+}
 /**
  * Decide whether `tool` is allowed under an agent's tool contract. Pure, so the
  * rail is unit-testable without spawning anything.
@@ -114,20 +132,88 @@ function readActiveAgent(cwd) {
 // PreToolUse-hook decision
 // ---------------------------------------------------------------------------
 /**
- * PreToolUse-hook decision. If an agent is active, parse its compiled `.md`
- * tool contract and allow the call only when the tool is in the allowlist;
- * otherwise block and tell the model which tools it may use. With no active
- * agent (or an agent that inherits all tools), always allow — the rail only
- * constrains agents that declared a contract.
+ * EXPERIMENTAL — parked (P3), flat-only, do NOT auto-wire. This deterministic
+ * subagent-window tracking assumes FLAT dispatch (one active subagent at a time);
+ * it is NOT nesting-safe — Claude Code v2.1.172 added depth-5 nested subagents, so
+ * correct tracking needs a depth-aware STACK (push on dispatch, pop on SubagentStop)
+ * + verifying the spawn tool name (`Agent` vs `Task`). See
+ * research/effect-boundary-design.md ("Why dropped").
+ *
+ * Resolve a `Task` tool's `subagent_type` to the compiled agent `.md` to
+ * activate, or null when none is found. The DETERMINISTIC open signal that
+ * replaces the model-invoked `agent-start`: Claude Code fires `PreToolUse` for
+ * the parent's `Task` dispatch (and `SubagentStop` when it returns), so the
+ * harness — not the model — brackets the subagent's active window. The name is
+ * the last ":"-segment (a `--plugin-dir` subagent_type is namespaced
+ * "plugin:name"); searched in `agents/` under the cwd then the plugin root. A
+ * path under the cwd is returned relative (readActiveAgent resolves vs cwd); a
+ * plugin-root hit is absolute. Returns null on an unknown agent (fail-open: an
+ * unresolved subagent is simply not gated, exactly as before agent-start ran).
+ */
+function resolveDispatchedAgent(subagentType, cwd, pluginRoot) {
+    const name = subagentType.split(":").pop()?.trim();
+    if (!name)
+        return null;
+    const rel = (0, node_path_1.join)("agents", `${name}.md`);
+    if ((0, node_fs_1.existsSync)((0, node_path_1.resolve)(cwd, rel)))
+        return rel;
+    if (pluginRoot && (0, node_fs_1.existsSync)((0, node_path_1.resolve)(pluginRoot, rel))) {
+        return (0, node_path_1.resolve)(pluginRoot, rel);
+    }
+    return null;
+}
+/**
+ * The agent `.md` to activate for a `PreToolUse(Task)` event, or null. Pure
+ * (reads `tool_input.subagent_type`, resolves via {@link resolveDispatchedAgent}).
  */
-function evaluatePreToolUse(cwd, tool) {
+function decideTaskDispatch(toolInput, cwd, pluginRoot) {
+    const st = toolInput?.subagent_type;
+    if (typeof st !== "string" || !st)
+        return null;
+    return resolveDispatchedAgent(st, cwd, pluginRoot);
+}
+/**
+ * PreToolUse-hook decision. If an agent is active, enforce BOTH deterministic
+ * rails its compiled `.md` declares, in order:
+ *
+ *  1. the tool-contract rail (`tools:`) — allow only listed tools;
+ *  2. the purity gate (`vigiles:purity:`) — allow only calls within the declared
+ *     effect floor, refining `Bash` by the live `command` (`decidePurityGate`).
+ *
+ * The first to deny wins, feeding its reason back to the model. With no active
+ * agent (or one that declared neither contract), always allow — the rails only
+ * constrain agents that opted in.
+ */
+function evaluatePreToolUse(cwd, tool, command) {
     const agentPath = readActiveAgent(cwd);
     if (!agentPath)
         return { allow: true, message: "" };
     const full = (0, node_path_1.resolve)(cwd, agentPath);
     if (!(0, node_fs_1.existsSync)(full))
         return { allow: true, message: "" };
-    const allowed = parseAgentTools((0, node_fs_1.readFileSync)(full, "utf-8"));
-    return decidePreToolUse(allowed, tool);
+    const md = (0, node_fs_1.readFileSync)(full, "utf-8");
+    // 1) Tool-contract rail — the declared allowlist.
+    const rail = decidePreToolUse(parseAgentTools(md), tool);
+    if (!rail.allow)
+        return rail;
+    // 2) Purity gate — the declared effect floor, refined by the live command.
+    //    If an effect boundary is declared, tighten to "pure" outside it and apply
+    //    the declared purity (or "unrestricted") inside.
+    const purity = parseAgentPurity(md);
+    const boundary = (0, effect_region_js_1.hasEffectBoundary)(md);
+    if (boundary) {
+        const effective = (0, effect_region_js_1.readEffectActive)(cwd)
+            ? (purity ?? "unrestricted")
+            : "pure";
+        const gate = (0, effects_js_1.decidePurityGate)(effective, tool, command, dialect_js_1.claudeCodeDialect);
+        if (!gate.allow)
+            return gate;
+    }
+    else if (purity) {
+        const gate = (0, effects_js_1.decidePurityGate)(purity, tool, command, dialect_js_1.claudeCodeDialect);
+        if (!gate.allow)
+            return gate;
+    }
+    return { allow: true, message: "" };
 }
 //# sourceMappingURL=agent-runtime.js.map

package/dist/adapters/claude-code/dialect.d.ts

@@ -8,6 +8,40 @@
  * in its adapter (e.g. `src/adapters/codex/dialect.ts` exporting `codexDialect`).
  */
 import type { HarnessDialect } from "../../core/dialect.js";
+/**
+ * The Claude Code built-in subagent tool catalog as a `const` tuple, so a typed
+ * authoring surface can derive a LITERAL union (`ClaudeCodeBuiltinTool`) from it.
+ * `claudeCodeDialect.builtinAgentTools` references this same array — one source
+ * of truth for the runtime catalog AND the compile-time tool vocabulary.
+ */
+export declare const claudeCodeBuiltinAgentTools: readonly ["Read", "Write", "Edit", "MultiEdit", "Bash", "BashOutput", "KillBash", "Grep", "Glob", "LS", "WebSearch", "WebFetch", "NotebookEdit", "TodoWrite", "Task", "Skill"];
+/**
+ * The Claude Code side-effecting tools as a `const` tuple (the complement of
+ * read-only within `builtinAgentTools`). `claudeCodeDialect.sideEffectingTools`
+ * references this; the typed vocabulary derives the read-only / bounded splits.
+ */
+export declare const claudeCodeSideEffectingTools: readonly ["Bash", "BashOutput", "KillBash", "Edit", "MultiEdit", "Write", "NotebookEdit", "WebFetch", "WebSearch", "Skill", "Task", "TodoWrite"];
 export declare const claudeCodeDialect: HarnessDialect;
+/** Every Claude Code built-in subagent tool (literal union). */
+export type ClaudeCodeBuiltinTool = (typeof claudeCodeBuiltinAgentTools)[number];
+/** The side-effecting subset (literal union). */
+export type ClaudeCodeSideEffectingTool = (typeof claudeCodeSideEffectingTools)[number];
+/**
+ * Read-only built-in tools — the complement of the side-effecting set within
+ * the built-in catalog. The tools a `pure` CC unit may declare.
+ */
+export type ClaudeCodeReadOnlyTool = Exclude<ClaudeCodeBuiltinTool, ClaudeCodeSideEffectingTool>;
+/**
+ * Tools a `bounded` CC unit may declare: read-only ∪ the decidable
+ * side-effecting tools (Write/Edit/MultiEdit/NotebookEdit) ∪ `Bash` (its
+ * command is decided at RUNTIME by the gate). Bars MCP / unknown / wildcard —
+ * those are simply not in the built-in union, so listing one is a `tsc` error.
+ *
+ * NOTE: `BashOutput`/`KillBash` are read-only-ish helpers tied to a running
+ * Bash; `Bash` is the admitting tool, so they're included via the read-only
+ * exclusion path only if read-only — here they stay side-effecting, hence the
+ * explicit add of the bounded-decidable set plus `Bash`.
+ */
+export type ClaudeCodeBoundedTool = ClaudeCodeReadOnlyTool | "Write" | "Edit" | "MultiEdit" | "NotebookEdit" | "Bash";
 export type { HarnessDialect } from "../../core/dialect.js";
 //# sourceMappingURL=dialect.d.ts.map

package/dist/adapters/claude-code/dialect.js

@@ -1,28 +1,54 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.claudeCodeDialect = void 0;
+exports.claudeCodeDialect = exports.claudeCodeSideEffectingTools = exports.claudeCodeBuiltinAgentTools = void 0;
+/**
+ * The Claude Code built-in subagent tool catalog as a `const` tuple, so a typed
+ * authoring surface can derive a LITERAL union (`ClaudeCodeBuiltinTool`) from it.
+ * `claudeCodeDialect.builtinAgentTools` references this same array — one source
+ * of truth for the runtime catalog AND the compile-time tool vocabulary.
+ */
+exports.claudeCodeBuiltinAgentTools = [
+    "Read",
+    "Write",
+    "Edit",
+    "MultiEdit",
+    "Bash",
+    "BashOutput",
+    "KillBash",
+    "Grep",
+    "Glob",
+    "LS",
+    "WebSearch",
+    "WebFetch",
+    "NotebookEdit",
+    "TodoWrite",
+    "Task",
+    "Skill",
+];
+/**
+ * The Claude Code side-effecting tools as a `const` tuple (the complement of
+ * read-only within `builtinAgentTools`). `claudeCodeDialect.sideEffectingTools`
+ * references this; the typed vocabulary derives the read-only / bounded splits.
+ */
+exports.claudeCodeSideEffectingTools = [
+    "Bash",
+    "BashOutput",
+    "KillBash",
+    "Edit",
+    "MultiEdit",
+    "Write",
+    "NotebookEdit",
+    "WebFetch",
+    "WebSearch",
+    "Skill",
+    "Task",
+    "TodoWrite",
+];
 exports.claudeCodeDialect = {
     name: "claude-code",
     // The tool contract a subagent may declare — the rails it runs on. Anything
     // else must be an MCP tool, else it's a typo / nonexistent tool.
-    builtinAgentTools: [
-        "Read",
-        "Write",
-        "Edit",
-        "MultiEdit",
-        "Bash",
-        "BashOutput",
-        "KillBash",
-        "Grep",
-        "Glob",
-        "LS",
-        "WebSearch",
-        "WebFetch",
-        "NotebookEdit",
-        "TodoWrite",
-        "Task",
-        "Skill",
-    ],
+    builtinAgentTools: exports.claudeCodeBuiltinAgentTools,
     // Tools the platform never exposes to a subagent, whatever the list says — so
     // a subagent listing one is a guaranteed-dead reference only a compiler catches.
     neverAvailableTools: [
@@ -63,5 +89,11 @@ exports.claudeCodeDialect = {
     // Claude Code reads the full SKILL.md frontmatter set (description,
     // disable-model-invocation, argument-hint, …).
     skillFrontmatter: "claude-code",
+    // Tools that produce side effects in Claude Code. The complement — the
+    // read-only tools — are: Read, Grep, Glob, LS, ToolSearch (and LSP/Agent
+    // which are not in the subagent catalog). Bash is side-effecting because
+    // `cat` and `rm -rf` are the same tool at the tool-name level — the
+    // sandbox is the only closure for subprocess effects.
+    sideEffectingTools: exports.claudeCodeSideEffectingTools,
 };
 //# sourceMappingURL=dialect.js.map

package/dist/adapters/claude-code/effect-region.d.ts

@@ -0,0 +1,9 @@
+/** Record that the agent has entered an effect boundary. */
+export declare function setEffectActive(cwd: string): void;
+/** Clear the effect-active marker (the agent exited the effect boundary). */
+export declare function clearEffectActive(cwd: string): void;
+/** True iff the agent is currently inside an effect boundary. Tolerates malformed file. */
+export declare function readEffectActive(cwd: string): boolean;
+/** True iff the compiled markdown declares an effect boundary (`<!-- vigiles:effect -->`). */
+export declare function hasEffectBoundary(markdown: string): boolean;
+//# sourceMappingURL=effect-region.d.ts.map

package/dist/adapters/claude-code/effect-region.js

@@ -0,0 +1,45 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.setEffectActive = setEffectActive;
+exports.clearEffectActive = clearEffectActive;
+exports.readEffectActive = readEffectActive;
+exports.hasEffectBoundary = hasEffectBoundary;
+/**
+ * Effect-boundary position-aware state tracking.
+ * Mirrors active-unit tracking (.vigiles/active-agent.json / .vigiles/active-skill.json).
+ * The PreToolUse hook reads this file to decide whether the agent is inside an
+ * effect boundary (set by `vigiles effect-enter`) or outside it.
+ */
+const node_fs_1 = require("node:fs");
+const node_path_1 = require("node:path");
+const EFFECT_ACTIVE_PATH = ".vigiles/effect-active.json";
+/** Record that the agent has entered an effect boundary. */
+function setEffectActive(cwd) {
+    const p = (0, node_path_1.resolve)(cwd, EFFECT_ACTIVE_PATH);
+    (0, node_fs_1.mkdirSync)((0, node_path_1.dirname)(p), { recursive: true });
+    (0, node_fs_1.writeFileSync)(p, JSON.stringify({ active: true }) + "\n");
+}
+/** Clear the effect-active marker (the agent exited the effect boundary). */
+function clearEffectActive(cwd) {
+    const p = (0, node_path_1.resolve)(cwd, EFFECT_ACTIVE_PATH);
+    if ((0, node_fs_1.existsSync)(p))
+        (0, node_fs_1.rmSync)(p);
+}
+/** True iff the agent is currently inside an effect boundary. Tolerates malformed file. */
+function readEffectActive(cwd) {
+    const p = (0, node_path_1.resolve)(cwd, EFFECT_ACTIVE_PATH);
+    if (!(0, node_fs_1.existsSync)(p))
+        return false;
+    try {
+        const parsed = JSON.parse((0, node_fs_1.readFileSync)(p, "utf-8"));
+        return parsed.active === true;
+    }
+    catch {
+        return false;
+    }
+}
+/** True iff the compiled markdown declares an effect boundary (`<!-- vigiles:effect -->`). */
+function hasEffectBoundary(markdown) {
+    return markdown.includes("<!-- vigiles:effect -->");
+}
+//# sourceMappingURL=effect-region.js.map

package/dist/adapters/claude-code/layout.js

@@ -9,6 +9,9 @@ exports.claudeCodeLayout = {
     settingsFormat: "json",
     instructionFile: "CLAUDE.md",
     surfaceDirs: ["skills", "agents", "commands"],
+    skillDir: "skills",
+    agentDir: "agents",
+    commandDir: "commands",
     materializeRoot: ".claude",
     pluginRootToken: "${CLAUDE_PLUGIN_ROOT}",
     mcpConfigFile: ".mcp.json",

package/dist/adapters/claude-code/skill-runtime.d.ts

@@ -20,6 +20,7 @@
  * (`vigiles run-skill`). It is not a silent hook and runs nothing the spec
  * didn't declare as a gate.
  */
+import type { PurityLevel } from "../../core/effects.js";
 export type RuntimeGate = {
     readonly kind: "cmd";
     readonly command: string;
@@ -86,6 +87,30 @@ export declare function setActiveSkill(cwd: string, skillPath: string): void;
 export declare function clearActiveSkill(cwd: string): void;
 /** The path of the active skill, or null when none is in progress. */
 export declare function readActiveSkill(cwd: string): string | null;
+/**
+ * Parse the declared purity floor from a compiled SKILL.md — the
+ * `<!-- vigiles:purity:LEVEL -->` marker `compileSkill` emits (see
+ * `purityMarker` in compile.ts). Returns null when no marker is present (the
+ * skill declared no floor, so the purity gate imposes no constraint). Mirrors
+ * `parseAgentPurity` in agent-runtime.ts.
+ */
+export declare function parseSkillPurity(markdown: string): PurityLevel | null;
+/** A runtime allow/deny decision (mirrors PreToolDecision in agent-runtime.ts). */
+export interface SkillPreToolDecision {
+    /** Whether the tool call is allowed (true) or blocked (false). */
+    readonly allow: boolean;
+    /** Message fed back to the model on a block; empty on allow. */
+    readonly message: string;
+}
+/**
+ * PreToolUse purity gate for skills. Mirrors `evaluatePreToolUse` in
+ * agent-runtime.ts, but enforces ONLY the purity floor — skills have no
+ * tools-allowlist rail (that's a separate, future concern). If a skill is
+ * active and declares a `vigiles:purity:` marker, `decidePurityGate` checks
+ * the live call (refining `Bash` by the concrete command via `isReadOnlyBash`).
+ * With no active skill, a missing `.md`, or no purity marker, always allows.
+ */
+export declare function evaluateSkillPreToolUse(cwd: string, tool: string, command?: string): SkillPreToolDecision;
 export interface StopDecision {
     /** Whether the agent may stop (true) or must keep working (false). */
     readonly allow: boolean;

package/dist/adapters/claude-code/skill-runtime.js

@@ -30,10 +30,14 @@ exports.runSkillGates = runSkillGates;
 exports.setActiveSkill = setActiveSkill;
 exports.clearActiveSkill = clearActiveSkill;
 exports.readActiveSkill = readActiveSkill;
+exports.parseSkillPurity = parseSkillPurity;
+exports.evaluateSkillPreToolUse = evaluateSkillPreToolUse;
 exports.evaluateStopHook = evaluateStopHook;
 const node_child_process_1 = require("node:child_process");
 const node_fs_1 = require("node:fs");
 const node_path_1 = require("node:path");
+const effects_js_1 = require("../../core/effects.js");
+const dialect_js_1 = require("./dialect.js");
 const STEP_RE = /^###\s+Step\s+(\d+)/;
 const GATE_CMD_RE = /<!--\s*vigiles:gate\s+"([^"]*)"(?:\s+retry:(\d+))?\s*-->/;
 const GATE_FILE_RE = /<!--\s*vigiles:gate\s+file:(\S+)\s*-->/;
@@ -257,6 +261,42 @@ function readActiveSkill(cwd) {
         return null;
     }
 }
+// ---------------------------------------------------------------------------
+// Skill purity gate (mirrors agent-runtime.ts parseAgentPurity / evaluatePreToolUse)
+// ---------------------------------------------------------------------------
+const PURITY_RE = /<!--\s*vigiles:purity:(pure|bounded|unrestricted)\s*-->/;
+/**
+ * Parse the declared purity floor from a compiled SKILL.md — the
+ * `<!-- vigiles:purity:LEVEL -->` marker `compileSkill` emits (see
+ * `purityMarker` in compile.ts). Returns null when no marker is present (the
+ * skill declared no floor, so the purity gate imposes no constraint). Mirrors
+ * `parseAgentPurity` in agent-runtime.ts.
+ */
+function parseSkillPurity(markdown) {
+    const m = PURITY_RE.exec(markdown);
+    return m ? m[1] : null;
+}
+/**
+ * PreToolUse purity gate for skills. Mirrors `evaluatePreToolUse` in
+ * agent-runtime.ts, but enforces ONLY the purity floor — skills have no
+ * tools-allowlist rail (that's a separate, future concern). If a skill is
+ * active and declares a `vigiles:purity:` marker, `decidePurityGate` checks
+ * the live call (refining `Bash` by the concrete command via `isReadOnlyBash`).
+ * With no active skill, a missing `.md`, or no purity marker, always allows.
+ */
+function evaluateSkillPreToolUse(cwd, tool, command) {
+    const skillPath = readActiveSkill(cwd);
+    if (!skillPath)
+        return { allow: true, message: "" };
+    const full = (0, node_path_1.resolve)(cwd, skillPath);
+    if (!(0, node_fs_1.existsSync)(full))
+        return { allow: true, message: "" };
+    const md = (0, node_fs_1.readFileSync)(full, "utf-8");
+    const purity = parseSkillPurity(md);
+    if (!purity)
+        return { allow: true, message: "" };
+    return (0, effects_js_1.decidePurityGate)(purity, tool, command, dialect_js_1.claudeCodeDialect);
+}
 /**
  * Stop-hook decision. If a skill is active and declares a result gate, run it:
  * allow the stop only when the gate passes; otherwise block and tell the model