name: skill-builder description: "Create, audit, optimize Claude Code skills. Commands: skills, list, new, strip, optimize, agents, hooks, verify, inline, ledger, cascade, checksums, convert, shell-safety, route" when_to_use: "When creating, auditing, or optimizing Claude Code skills, or when working with SKILL.md files, hooks, or agents" argument-hint: "[command] [skill] [--execute]" version: "1.5" minimum-effort-level: high allowed-tools: Read, Glob, Grep, Write, Edit, TaskCreate, TaskUpdate, TaskList, TaskGet hooks: PostCompact: - hooks: - type: command command: "echo '{"additionalContext": "REMINDER: Directives are sacred. Never reword, paraphrase, or summarize text between  markers. Optimization is restructuring, not rewriting. Move content — never rewrite it."}'" statusMessage: "Re-injecting directive awareness..."

Skill Builder

Quick Commands

Command	Action
`/skill-builder`	Full audit: runs optimize + agents + hooks in display mode for all skills
`/skill-builder audit`	Same as above
`/skill-builder audit --quick`	Lightweight audit: frontmatter + line counts + priority fixes only
`/skill-builder cascade [skill]`	Validation cascade analysis: detect over-validation suppressing output
`/skill-builder dev [command]`	Run any command with skill-builder itself included

Directives

"When a decision needs to be made that isn't overtly obvious, and guesses are involved, AGENTS ARE MANDATORY, in order to provide additional input in decision making."

— Added 2026-02-22, source: user directive

"Each agent being created by this system always has to have an appropriate persona that is not being used anywhere else."

— Added 2026-02-22, source: user directive

"When deploying a Team, one of the team member's persona is a research assistant who will research the issue using read-only reference tools. Other team members may also make requests from the research assistant to help augment the outcome."

— Added 2026-02-23, source: user directive (tool specifics in references/agents-teams.md)

"When the dev flag gets called, you ALWAYS concentrate on the distribution files first, then sync changes to the .claude directory after."

— Added 2026-05-08, source: user directive (after dev edits repeatedly landed in the runtime copy instead of the source distribution)

"No hooks! We don't distribute hooks. The project only makes hooks on the host system."

— Added 2026-05-08, source: user directive

"Exception to the no-hooks-distribution rule: skill-builder's own load-bearing enforcement hooks — protect-directives.sh and unique-persona.sh — DO ship in the source distribution and the installer fetches them. They protect two sacred user directives (no rewording of immutable blocks; persona uniqueness across agents). Without them, every fresh install silently loses load-bearing enforcement. The general no-distribute rule still applies to every other hook on the host system. Wiring into settings.local.json remains host-local."

— Added 2026-05-11, source: user directive (after the regenerate-and-rewire loop revealed that no-distribute leaves load-bearing enforcement off on every fresh host).

CHECKPOINT — Non-Obvious Decision Gate:

Before committing to any classification, structural change, or content-removal decision that depends on interpretation, list the decision's alternatives in one sentence each.
IF exactly one alternative matches a concrete, measurable criterion (ID match, regex match, frontmatter field present/absent, file exists/absent) → CONTINUE without an agent.
IF two or more alternatives are plausible AND the selection requires judgment on wording, scope, priority, or fit → STOP. Spawn at least one agent via the Task tool (or an agent panel per the relevant procedure — e.g., optimize.md § 4b, § 5b) to supply independent input.
Read the agent findings. Where agents agree → proceed with the agreed alternative. Where agents disagree → default to the safer/conservative alternative.
IF an agent was required but skipped → STOP. Report to user: "Agent consultation skipped for a non-obvious decision. Respawn with agent input before proceeding."

CHECKPOINT — Persona Assignment Gate:

Before writing or editing any AGENT.md file, extract the proposed persona string from the agent frontmatter.
Read references/agents-personas.md § "Persona assignment rules" to confirm the persona fits the agent's stated role (task, scope, perspective).
Glob BOTH agent-file forms and read each file's persona field: .claude/skills/*/agents/*.md (flat-file agents like agents/failure-triage.md) AND .claude/skills/*/agents/*/AGENT.md (subdirectory-form agents like agents/optimize-diff-auditor/AGENT.md). Union the two globs — agents may live in either form, and using only one form silently drops uniqueness coverage for the other half of the population.
IF the proposed persona string matches any existing persona verbatim OR paraphrases one already in use (same core identity, different words) → STOP. Report: "Persona conflicts with [path]: '[existing persona]'. Choose a different persona."
IF no duplicate AND the persona fits the role (step 2 passed) → CONTINUE to write the AGENT.md.
There is no shipped backstop hook for this gate (skill-builder does not distribute pre-built hook scripts; cross-platform compatibility takes precedence). The CHECKPOINT above IS the enforcement — follow it literally during authorship. Users who want a deterministic backstop on their own systems can generate one via /skill-builder hooks dev skill-builder --execute, which builds an OS-appropriate hook locally without shipping it.

CHECKPOINT — Team Research Assistant Gate:

Detect team deployment: the procedure invokes TeamCreate, uses language like "Spawn teammates", or explicitly assembles multiple parallel agents under one task.
IF detected → read references/agents-teams.md § "Individual vs. team routing" for the allowed research-assistant tool list (read-only reference tools).
Enumerate the planned team members and their personas. IF none are labeled as the research assistant → STOP. Add a research-assistant team member with the read-only reference toolset from step 2 before spawning.
IF a research assistant is present → confirm its allowed-tools contains ONLY read-only reference tools. Strip any write or side-effect tools from the research assistant before spawn.
After spawn, surface to other team members: "Research assistant '[persona]' is available for read-only reference queries. Route lookups through the assistant rather than duplicating research."
IF a team is deployed without a research assistant (step 3 skipped) → STOP. Report: "Team deployment blocked — no research assistant in team composition."

CHECKPOINT — Source-First Ordering Gate (fires when dev_mode == true):

Maintainer mode detection: does ${CLAUDE_PROJECT_DIR}/skill-builder/SKILL.md exist?
- YES → maintainer mode active. Continue.
- NO → end-user mode. This CHECKPOINT is a no-op.
Track per-session state: which skill-builder/<path> files have been edited so far in this session? Maintain a mental ledger of source paths touched.
Before issuing ANY Edit/Write tool call on a file under .claude/skills/skill-builder/<path> or skill-builder/<path>:
- IF the planned target is skill-builder/<path> → CONTINUE. This is the canonical first-pass edit.
- IF the planned target is .claude/skills/skill-builder/<path>:
  - IF skill-builder/<path> has already been edited in this session → CONTINUE. This is the mirror phase.
  - IF skill-builder/<path> has NOT been edited in this session → STOP. REWRITE the target to skill-builder/<path> BEFORE issuing the call. Edit source first.
Reverse-order forbidden: never edit runtime first then "sync back to source." The runtime contains intentional runtime-only content (local hooks frontmatter, sidecars) that must NOT propagate to source.
End-of-session check: git status --short -- skill-builder/. Empty when changes were expected = FAIL. Report: "Source-first ordering violated. Edits landed in the runtime only. Reverse order and retry."

CHECKPOINT — No-Distribute-Hooks Gate:

Define the EXCEPTION_HOOKS set: { protect-directives.sh, unique-persona.sh }. Every step below applies to all hooks EXCEPT those in this set; the exception steps (1b, 3b) cover the two named hooks explicitly.
Before adding any hook script under skill-builder/hooks/ whose basename is NOT in EXCEPTION_HOOKS → STOP. The source distribution MUST NOT contain hook scripts other than the two named exceptions. Hooks live only in the runtime copy on the host system.
- 2b. Exception path: adding skill-builder/hooks/protect-directives.sh or skill-builder/hooks/unique-persona.sh is PERMITTED and REQUIRED per the 2026-05-11 sacred directive. These ship in source.
Before adding a hooks: frontmatter block to source skill-builder/SKILL.md → STOP. Source SKILL.md MUST NOT declare hooks. The runtime SKILL.md may declare hooks the host has generated locally; source must not.
Before adding any line to the install script that fetches a hook script via curl → check against EXCEPTION_HOOKS.
- 4a. If the basename is in EXCEPTION_HOOKS → PERMITTED. The installer is expected to fetch these two files. Confirm the fetch loop targets .claude/skills/skill-builder/hooks/ on the host.
- 4b. If the basename is NOT in EXCEPTION_HOOKS → STOP. Adding the fetch line violates the directive.
Hooks ARE permitted in the runtime copy (.claude/skills/skill-builder/hooks/) and in runtime SKILL.md frontmatter, but only when generated on the host system via /skill-builder hooks <skill> --execute or maintained by hand by the host operator. The two EXCEPTION_HOOKS additionally arrive via the installer's fetch loop. Runtime hooks NOT in EXCEPTION_HOOKS never propagate back to the source distribution.
IF a workflow proposes shipping a hook NOT in EXCEPTION_HOOKS via install, adding non-exception hook scripts to skill-builder/, or declaring hooks in source frontmatter → REFUSE and report: "No-distribute-hooks directive violated. Hooks are made on the host system only — only protect-directives.sh and unique-persona.sh are permitted in source per the 2026-05-11 exception."

Phase 0: Dev Path Discipline (BLOCKING — maintainer mode)

When dev_mode == true AND ${CLAUDE_PROJECT_DIR}/skill-builder/SKILL.md exists (this repo IS the skill-builder source distribution), every Edit/Write on a skill-builder file MUST target the source path under skill-builder/... BEFORE any mirror to the runtime copy at .claude/skills/skill-builder/....

The runtime is gitignored. It gets overwritten on every bash install. Runtime-only edits never reach end users.

Mandatory order — non-negotiable:

Edit skill-builder/<path> first (the source distribution under repo root).
Then mirror the same change to .claude/skills/skill-builder/<path> so the running session matches source. The mirror is a content sync, not a wholesale overwrite. Preserve runtime-only content the source intentionally lacks: local hooks frontmatter, .directives.sha sidecars, generated artifacts.
Never reverse the order. Runtime contains intentional runtime-only content that must NOT propagate to source.

Hooks-in-source exception: Two skill-builder hooks ship in the source distribution per the 2026-05-11 sacred directive — protect-directives.sh and unique-persona.sh. When dev mode targets either file, the canonical source path is skill-builder/hooks/<name> — edit there first, then mirror to .claude/skills/skill-builder/hooks/<name>. Every other hook file remains runtime-only and follows the original no-distribute rule.

CHECKPOINT — fires before any skill-builder Read/Edit/Write when dev is in the invocation:

Maintainer mode: does ${CLAUDE_PROJECT_DIR}/skill-builder/SKILL.md exist?
- YES → maintainer mode active. Continue.
- NO → end-user mode. This CHECKPOINT is a no-op. Proceed to dispatch.
For every planned Edit/Write whose path starts with .claude/skills/skill-builder/:
- REWRITE the path BEFORE issuing the tool call: replace .claude/skills/skill-builder/ with skill-builder/. The source path is the canonical first-pass edit target.
- IF the source file does not exist while the runtime file does → STOP. Report: "Runtime is ahead of source for [path]. Determine canonical state before editing." Do not auto-mirror.
- Hook path exception: paths matching .claude/skills/skill-builder/hooks/protect-directives.sh or .claude/skills/skill-builder/hooks/unique-persona.sh rewrite to skill-builder/hooks/<name> (these two ship in source per the 2026-05-11 directive). Any OTHER .claude/skills/skill-builder/hooks/*.sh path stays runtime-only — do NOT rewrite to source for those.
For Reads on skill-builder content: prefer skill-builder/<path> so planning grounds on canonical source. Reads from runtime are allowed but second choice — the runtime may be stale.
After all source edits land, perform the runtime mirror as a separate, explicit phase. For each skill-builder/<path> modified in this session, replicate the same change to .claude/skills/skill-builder/<path>. Touch only the changed sections; do not overwrite runtime-only frontmatter, hook scripts, or sidecars.
End-of-session check: git status --short -- skill-builder/. Empty output when changes were expected = FAIL. The edits landed in the runtime only. Reverse order and retry from step 2.

No hook backstop for this gate. Per the user directive "No hooks! We don't distribute hooks. The project only makes hooks on the host system" (see § Directives), there is no shipped hook for Phase 0 enforcement. The CHECKPOINT above IS the enforcement and must run during authorship. A maintainer who wants a local mechanical backstop on their own host can generate one via /skill-builder hooks dev skill-builder --execute, but that is a host-system action and is never part of the source distribution.

Commands

All commands operate in display mode by default. Add --execute to apply changes. Before executing any command, read its procedure file from references/procedures/.

Command	Procedure	Summary
`audit`	audit.md	Full system audit
`audit --quick`	audit.md	Lightweight: frontmatter + line counts
`cascade [skill]`	cascade.md	Validation cascade analysis (diagnostic only)
`convert [skill]`	convert.md	Convert 4.6-era skill to 4.7-compatible (annotations + explicit steps)
`optimize [skill]`	optimize.md	Restructure for context efficiency
`optimize claude.md`	claude-md.md	Extract domain content to skills
`agents [skill]`	agents.md	Analyze/create agents
`hooks [skill]`	hooks.md	Inventory/create hooks
`new [name]`	new.md	Create skill from template
`strip [skill]`	strip.md	Delete a skill and remove all cross-references
`inline [skill] [directive]`	inline.md	Quick-add directive
`skills`	skills.md	List local skills
`list [skill]`	list.md	Show modes/options
`verify`	verify.md	Health check (headless-compatible)
`ledger`	ledger.md	Create Awareness Ledger
`checksums [skill]`	checksums.md	Generate/verify directive checksums
`shell-safety [mode] [path]`	shell-safety.md	Write / audit / lint shell code and JSON-embedded shell for pitfalls
`route [mode]`	route.md	Maintain `/route` skill index and embed route-consultation hooks into other skills
`update`	(inline below)	Update to latest version

The `update` Command

Re-run the installer to update skill-builder to the latest version.

The installer issues many file writes and bash calls in sequence. Without auto-accept, the user will be prompted to approve each one. Claude Code does NOT expose a way for a skill to flip the session into "accept edits on" mode programmatically, nor to detect the current permission mode at runtime — mode changes require the user to press Shift+Tab. The procedure below therefore prompts the user to enable auto-accept before the installer runs.

CHECKPOINT — fires when /skill-builder update is invoked:

BEFORE running the installer, output this notice to the user verbatim and STOP for their acknowledgement:

Before I run the installer, please enable "accept edits on" mode so you don't get prompted for every file write and bash call.

Press Shift+Tab until the prompt indicator shows "accept edits on" (it cycles: default → accept edits on → plan mode).

I cannot detect or set this mode from inside the session — it has to be you. Reply with anything (e.g., "go") once it's enabled and I'll run the installer.
After the user acknowledges, run the installer via Bash: bash -c "$(curl -fsSL https://raw.githubusercontent.com/odysseyalive/claude-enforcer/main/install)"
Tell the user: "Restart Claude Code to load the updated skill." The current session still has the old skill loaded in memory, so start a new conversation. Once you're back, run /skill-builder audit — updates often add new recommendations that apply to your existing skills.

The `convert` Command

Convert existing Opus 4.6-era skills to Opus 4.7-compatible execution. User directives stay verbatim and receive enforcement annotations (machine-generated CHECKPOINT blocks beneath the sacred block); skill-builder machinery (workflow steps, grounding statements) is rewritten in-place for literal execution.

Display mode (default): /skill-builder convert [skill] — report what would change
Execute mode: /skill-builder convert [skill] --execute — apply changes
Batch display: /skill-builder convert --all — summary across all skills
Batch execute: /skill-builder convert --all --execute — convert every skill in sequence (one task per skill; the task list survives context compaction)

High-risk command — defaults to display mode, requires --execute to modify files.

Grounding: Read references/procedures/convert.md for the full procedure, references/templates.md § "Enforcement Annotation Template" for the annotation format, and references/enforcement.md § "Opus 4.7 Behavioral Contract" for the literal-execution model.

The `route` Command

Maintain the /route skill — a glorified, auto-generated index of every installed skill — and embed route-consultation hooks into other skills so the AI dispatches through registered skills instead of freelancing. /route is a peer to intent-router, not a replacement: intent-router handles freeform /skill-builder <text> invocations within skill-builder; /route handles user-level task routing across every installed skill.

/skill-builder route index — scan all skills, regenerate /route's catalog. Bootstraps /route if missing. Diffs against the prior index and reports NEW / REMOVED / UPDATED / UNCHANGED. Default mode is execute (low-risk; only writes auto-generated content inside /route).
/skill-builder route index --dry-run — display-only summary of what would change.
/skill-builder route embed — display mode (high-risk default). For each skill, classify NEW / REFRESH / REMOVE / NOOP based on workflow heuristics + a reconciliation against any embed blocks already on disk.
/skill-builder route embed --execute — apply the planned embeds, refreshes, and removals; auto-runs route index --execute afterward to keep the catalog current.
/skill-builder route embed --remove [skill] — manually opt a skill out of the route gate.

Both index and embed are intelligent on re-run: index diffs against the prior catalog and rewrites idempotently; embed reconciles against existing  markers and either refreshes them, removes them when the skill no longer qualifies, or adds them where workflows now require routing.

Audit integration: route index is appended as the second-to-last item in audit's task list, and route embed is the last item. See audit.md § Step 4g.

Grounding: Read references/procedures/route.md for the full procedure, including the embed block format, reconciliation rules, and the /route skill template used during bootstrap.

The `strip` Command

Delete a skill completely and remove every connection to it from other skills, settings, hook bindings, and dev-repo manifests. The destructive counterpart to new.

Display mode (default): /skill-builder strip [skill] — produce an impact report listing every file to be deleted, every cross-reference to be removed, dependent skills, and BREAKING status if any HARD references exist
Execute mode: /skill-builder strip [skill] --execute — apply the deletion plan
Breaking confirmation: /skill-builder strip [skill] --execute --confirm-breaking — required when the target has HARD references in other skills (workflow Read instructions, hook scripts, or AGENT.md grounding)

Destructive command — defaults to display mode, requires --execute. Stripping skill-builder itself is HARD-REFUSED even with the dev prefix; the prefix permits self-modification, not self-deletion.

After deletion, the procedure auto-runs route index --execute to drop the target from the /route catalog (when /route is installed).

Grounding: Read references/procedures/strip.md for the full procedure, including the 15 cross-reference detection patterns, dependent classification, settings.local.json mutation rules, and the strict task ordering (sweep references before deletion).

The `shell-safety` Command

Write, audit, and lint shell code (scripts, hook commands, JSON-embedded shell strings) against the canonical pitfall rule set. Used internally by hooks and verify, and available for direct user invocation.

Write: /skill-builder shell-safety write [target] — generate a new script or JSON shell entry from a safe-default template
Audit: /skill-builder shell-safety audit [path] — scan for pitfalls; with --execute, patch the mechanical-safe ones
Lint: /skill-builder shell-safety lint [file] — read-only single-file check (exit 1 on findings, composes with &&)

Grounding: Read references/procedures/shell-safety.md for the full procedure, references/shell-safety/rules.md for the pitfall catalog, references/shell-safety/templates.md for safe scaffolds, and references/shell-safety/audit-patterns.md for detection regexes.

Self-Exclusion Rule

The skill-builder skill MUST be excluded from all actions (audit, optimize, agents, hooks, skills list) unless the command is prefixed with dev.

/skill-builder audit → audits all skills EXCEPT skill-builder
/skill-builder optimize some-skill → works normally
/skill-builder optimize skill-builder → REFUSED. Say: "skill-builder is excluded from its own actions. Use dev prefix: /skill-builder dev optimize skill-builder"
/skill-builder dev audit → includes skill-builder in the audit
/skill-builder dev optimize skill-builder → allowed

Detection: If the first argument after the command is dev, strip it and proceed with self-inclusion enabled. Otherwise, skip any skill whose name is skill-builder when iterating skills, and refuse if skill-builder is explicitly named as a target.

CHECKPOINT — apply before dispatching to any procedure (Opus 4.7 literal-execution gate):

Parse the invocation. Is the first positional argument the literal string dev?
- YES → dev_mode = true; strip dev from the argument list; continue.
- NO → dev_mode = false.
Extract the first remaining positional argument as first_arg.
Define the known-command set: { audit, optimize, agents, hooks, new, inline, skills, list, verify, ledger, cascade, checksums, convert, shell-safety, route, update }.
IF first_arg is empty (no arguments remaining) → dispatch to the default full-audit flow per § Quick Commands. Do NOT invoke the intent router. STOP this CHECKPOINT.
IF first_arg is in the known-command set → treat it as the command name. Determine whether a skill target was specified in the remaining arguments. CONTINUE to step 7.
IF first_arg is NOT in the known-command set AND the remaining argument string is non-empty →
- STOP normal dispatch.
- Ground on references/procedures/intent-router.md (read-before-use).
- Invoke the Intent Router procedure, passing the full remaining argument string (including first_arg) as the freeform intent, and dev_mode as context.
- The router is responsible for either re-dispatching to a known command (in which case resume this CHECKPOINT at step 7 with the router's resolved command and target) or halting with an AskUserQuestion / explanatory message.
- The router NEVER modifies files; all file-touching work happens in the re-dispatched command.
IF dev_mode == false:
- IF the skill target is the literal string skill-builder → REFUSE and STOP. Print: "skill-builder is excluded from its own actions. Use dev prefix: /skill-builder dev [command] skill-builder". Do not dispatch.
- IF no target was specified (all-skills/iteration mode) → the skill set passed into the procedure MUST have skill-builder filtered out before any per-skill iteration begins.
IF dev_mode == true → skill-builder is included normally.
Dispatch to the procedure file.

This CHECKPOINT fires every invocation. Procedure files repeat it in their own preflight blocks for defense in depth — 4.7 executes each file literally, so both gates matter.

Dev Path Discipline (defense-in-depth pointer): Phase 0 at the top of this SKILL.md is the primary gate for source-vs-runtime path resolution in maintainer mode. It fires BEFORE this CHECKPOINT in document order. If Phase 0 ran cleanly, the path was already rewritten to the source location before reaching dispatch. If you reached this point and a planned Edit/Write still targets .claude/skills/skill-builder/..., STOP and re-read § Phase 0.

Post-dev check: After any dev command that modifies skill-builder files, run BOTH of the following:

Manifest check. Glob skill-builder/**/*.md. Compare against the files downloaded in the installer's for ref, for proc, and for ss loops, plus the explicit curl lines for SKILL.md and agents/*/AGENT.md. Flag any new/renamed/removed files the installer doesn't handle. This prevents drift between the repo and what users receive on install. (skill-builder no longer ships hook scripts; users generate per-system hooks via /skill-builder hooks if they want them.)
Source-edit verification. Run git status --short -- skill-builder/. If the dev session was expected to produce changes and the output is empty, the edits landed in the runtime copy. Report the failure with the runtime/source diff so the user can mirror canonical changes back to source.

Display/Execute Mode Convention

Commands are classified by risk level, which determines their default mode:

Risk	Commands	Default Mode
Low-risk (additive, non-destructive)	`new`, `inline`, `skills`, `list`, `verify`, `ledger`, `checksums`, `route index`	Execute directly
High-risk (restructuring, modifying)	`optimize`, `agents`, `hooks`, `audit`, `cascade`, `convert`, `route embed`	Display mode (requires `--execute`)
Destructive (deletes files irreversibly)	`strip`	Display mode (requires `--execute`; `--confirm-breaking` if dependents exist)

Mode	Behavior	Flag
Display	Read-only plan of what would change	(default for high-risk)
Execute	Apply changes to files	`--execute` or (default for low-risk)

Rules

Low-risk commands execute immediately. new, inline, skills, list, verify, ledger, checksums, and route index do their work directly without requiring --execute. They are additive or read-only — there is nothing to preview. (route index is auto-generated content inside the /route skill only — idempotent regeneration.)
High-risk commands default to display mode. Running /skill-builder optimize my-skill shows what would change without modifying anything. Add --execute to apply.
Audit always calls sub-commands in display mode, then offers the user a choice of which to execute.
Execution requires a task plan. When a high-risk command runs with --execute, the command MUST:
- First produce a numbered task list using TaskCreate, one task per discrete action
- Execute each task sequentially, marking progress via TaskUpdate
- This ensures context can be refreshed mid-execution without losing track, no tasks get forgotten during long context windows, and the user can see progress and resume if interrupted
Scope discipline during execution. Execute ONLY the tasks in the task list. Do not add bonus tasks, expand scope, or create deliverables not in the original plan. If execution reveals a new opportunity, note it in the completion report — do not act on it. The task list is the contract.
Post-action chaining. Any action that modifies a skill (new, inline, adding directives) automatically chains into a scoped mini-audit for the affected skill — running optimize, agents, and hooks in display mode, then offering execution choices. Use --no-chain to suppress.

Core Principles

Read references/principles.md before running any high-risk command (optimize, agents, hooks, audit, cascade, convert). That file contains the full Core Principles, Sacred Directive Pattern, Output Discipline rules, and Grounding Protocol. It was split out of SKILL.md during the 4.7 upgrade to reduce always-loaded context weight.

IMPORTANT: Never break anything. Optimization is RESTRUCTURING, not REWRITING. MOVE content, don't rewrite it. PRESERVE all directives exactly. KEEP all workflows intact. If the original author reviewed the result, they should say "this does exactly what mine did, just organized differently."

Grounding

Grounding protocol (read-before-use, state which pattern will be used) is documented in references/principles.md § "Grounding Protocol".

Reference files:

references/principles.md — Core Principles, Sacred Directive Pattern, Output Discipline, Grounding Protocol (READ FIRST for high-risk commands)
references/enforcement.md — Hook JSON, permissions, context mutability, provenance permission model
references/agents.md — Agent templates, opportunity detection, creation workflow
references/agents-personas.md — Persona assignment rules, selection heuristic, research backing
references/agents-teams.md — Individual vs. team routing, invocation patterns, mandatory agent situations
references/templates.md — Skill directory layout, SKILL.md template, frontmatter
references/optimization-examples.md — Before/after examples, optimization targets
references/portability.md — Install instructions, rule-to-skill conversion
references/patterns.md — Lessons learned
references/platform.md — Claude Code skill platform architecture, frontmatter fields, listing budget, invocation flow
references/token-efficiency.md — Token-intensive pattern catalog and Token Efficiency Scan rules (optimize step 4e)
references/temporal-validation.md — Temporal risk classification, phrase mappings, hook generation spec
references/ledger-templates.md — Awareness Ledger record templates, agent definitions, consultation protocol
references/procedures/ — Per-command procedure files (audit, verify, optimize, agents, hooks, new, inline, ledger, cascade, checksums, shell-safety, etc.)
references/procedures/checksums.md — Directive checksum generation spec (scripts generated at runtime, not shipped)
references/procedures/shell-safety.md — Shell-safety subcommand procedure (write / audit / lint)
references/procedures/route.md — Route subcommand procedure (index + embed) with /route skill bootstrap template
references/shell-safety/ — Shell-safety rule set (rules.md, templates.md, audit-patterns.md) — the canonical pitfall catalog used by hooks, verify, and shell-safety
agents/optimize-diff-auditor/ — Post-optimize semantic equivalence verification agent

Name	skill-builder
Description	Create, audit, optimize Claude Code skills. Commands: skills, list, new, optimize, agents, hooks, verify, inline

skill-builder

SKILL.md

Skill Builder

Quick Commands

Directives

Phase 0: Dev Path Discipline (BLOCKING — maintainer mode)

Commands

The update Command

The convert Command

The route Command

The strip Command

The shell-safety Command

Self-Exclusion Rule

Display/Execute Mode Convention

Rules

Core Principles

Grounding

The `update` Command

The `convert` Command

The `route` Command

The `strip` Command

The `shell-safety` Command