Claudecode

name: {{skill}} description: {{description}}

model: haiku # haiku = lightweight read/report · omit = orchestration/reasoning

effort: high # high = multi-step orchestration · low = cheap scout · omit = default

allowed-tools: Read, Glob, Grep, Bash # explicit allowlist

argument-hint: "[issue-id]" # REQUIRED when the skill takes arguments — autocomplete hint

disable-model-invocation: true # user-triggered only (deploy, commit, …) — Claude never auto-invokes

user-invocable: false # background knowledge skill — Claude reads it, users can't invoke it

paths: src/**/*.ts # glob pattern(s) — activate only when working with matching files (drop the backticks)

disallowed-tools: Write, Edit # remove tools from the pool while the skill is active

context: fork # run the skill in a forked subagent instead of inline

agent: Explore # subagent to use when context: fork is set

{{skill}}

Rigid [gate type]. Match the user's language; keep technical identifiers unchanged.

Voice

Read ../../persona.md at the start of this skill. It is canonical for every user-facing string this skill emits, and its scope ends with the final report.

[IF the skill needs dynamic context — optional section, repo-wide convention]

Context

Auto-injected on Claude Code at skill load. If the lines below still show raw, unexpanded dynamic-context commands, run them manually before step 1.

• [Label]: ![read-only command]
• [Label]: ![read-only command]

[/ENDIF]

Workflow

1. Preconditions:
   • [List the things that must be true before this skill runs. Examples: MCP tools loaded, git repo verified, state file readable.]
2. [Step name]:
   • [Ordered actions. Use Bash / Read / MCP tools as needed. Keep bullets tight.]
3. [Step name]:
   • [...] N. [Final action — handoff, report, or stop]:
   • [...]

[IF hand-off menu] Present numbered options after the final action:

[voice intro line]
  (a) <label> → <what happens>
  (b) <label> → <what happens>
  (s) stop    → <clean exit message>

Branch on response. Exit skill when the chosen branch finishes. [/ENDIF]

[IF this skill hands to a specific next skill] Name the next skill explicitly, on its own line, so the chain cannot be guessed:

REQUIRED SUB-SKILL: Use <plugin>:<skill> [/ENDIF]

Final Report

{{plugin}}:{{skill}} report
  <Field>:        <value>
  <Field>:        <value>

Never

• Run git push, git commit, or git rebase.
• Mutate external services without explicit user confirmation.
• Skip the preconditions step.
• [Skill-specific don'ts]