run-tamirs-superpowers
Skill2 starsChanged 4 months agoMITView on GitHub
Use when running, validating, or testing the tamirs-superpowers Claude Code plugin — checks JSON validity, shellcheck hooks, SKILL.md frontmatter, statusline output, hook wiring, and absence of hardcoded paths or internal references. Trigger phrases: ''run the plugin health check'', ''validate the plugin'', ''test tamirs-superpowers'', ''smoke test the plugin'', ''check the plugin'', ''is the plugin clean''.
- Installs packages
In this repository46 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .cursor/rules/commit-conventions.mdcCursor rule
- .cursor/rules/cross-platform-handoff.mdcCursor rule
- .cursor/rules/dev-files-workspace.mdcCursor rule
- .cursor/rules/documentation-standards.mdcCursor rule
- .cursor/rules/gh-cli-preference.mdcCursor rule
- .cursor/rules/github-repository-policy.mdcCursor rule
- .cursor/rules/git-worktree-agent-workflow.mdcCursor rule
- .cursor/rules/hooks-guide.mdcCursor rule
- .cursor/rules/plugin-structure.mdcCursor rule
- .cursor/rules/plugin-version-bump.mdcCursor rule
- .cursor/rules/skill-frontmatter.mdcCursor rule
- .cursor/rules/skill-quality-standards.mdcCursor rule
- .cursor/rules/skills-guide.mdcCursor rule
- .cursor/rules/user-facing-script-standards.mdcCursor rule
- field-notebook-uiSkill
- diagnose-refusalSkill
- targeted-debugSkill
- decisionSkill
- deliver-devSkill
- diagnose-objectiveSkill
- orchestrate-devSkill
- plan-devSkill
- pr-devSkill
- start-devSkill
- switch-devSkill
- worker-devSkill
- changelog-reviewSkill
- dark-terminal-docSkill
- docs-reviewSkill
- platform-syncSkill
- mcp-builderSkill
- mcp-paginationSkill
- cleanupSkill
- github-policySkill
- multi-agent-repoSkill
- repo-scaffoldSkill
- repo-standardsSkill
- capture-configSkill
- find-skillSkill
- notify-setupSkill
- retroSkill
- session-reportSkill
- skill-creatorSkill
- usage-captureSkill
The file
Files
run-tamirs-superpowers
tamirs-superpowers is a Claude Code plugin — pure Markdown, JSON, and Bash. There is no build step. The plugin runs inside every Claude Code session automatically. The "interactive surface" is the plugin's validation harness (make validate) and the smoke.sh driver in this directory.
Prerequisites
brew install shellcheck jq # shellcheck for hook lint, jq for JSON checks
python3 -m pip install -q -r scripts/requirements-validate.txt # PyYAML for frontmatter validatorRun (agent path) — smoke.sh
The driver runs a 7-category health check. Always run from repo root:
bash .claude/skills/run-tamirs-superpowers/smoke.shWhat it checks (all verified in this session):
| Check | What it does |
1. make validate | shellcheck every tracked .sh at any depth, validate all JSON, validate SKILL.md frontmatter (portable core + metadata.tamirs + Claude extensions) |
2. scripts/statusline.sh | runs the live statusline and checks it produces output |
3. plugin.json | all required fields present, .statusLine is an object not a string |
| 4. SKILL.md quality | python3 scripts/validate-skill-frontmatter.py — tiered: portable core fails the build, metadata.tamirs warns while absent, Claude fields validated when present |
| 5. Hook wiring | every .sh in hooks/ is referenced in hooks.json |
6. No /Users/ paths | skills must use $CLAUDE_SKILL_DIR or relative paths, not hardcoded absolutes |
| 7. No employer/internal refs | internal-hostname shapes must be absent from skill files; no employer is named in the pattern — set $TAMIRS_EMPLOYER_PATTERN for a private one |
Exit 0 = all 7 sections pass. Exit 1 = at least one FAIL (not WARN).
Expected clean output:
PASS make validate
PASS statusline.sh — output: [unknown] ctx:--
PASS plugin.json .name = tamirs-superpowers
...
PASS: 31 | WARN: 4 | FAIL: 0
Plugin health: OK (warnings=4)Known warnings (not failures): none when hooks/ contains only lifecycle scripts wired in hooks.json.
scripts/github-mcp.sh is wired via .mcp.json, not hooks.json.
Run (human path) — make validate
For a quick syntax check without the full smoke test:
make validate # shellcheck + JSON + frontmatter
make lint # shellcheck onlyTo run the full Claude Code plugin validator (requires claude CLI):
make plugin-validateStatusline
The statusline script runs automatically in every Claude Code session. To test it in isolation:
bash scripts/statusline.shExpected output format: [<branch>] ctx:<n> | <model> | ...
Adding a skill
Skills live at skills/<domain>/<name>/SKILL.md. After adding or editing one:
make validate # must pass before committingFrontmatter fields required by the smoke test (in addition to name and description):
allowed-tools: [...]when_to_use: "..."(user-invocable skills only)metadata.updated-date: "YYYY-MM-DD"
Gotchas
statusLineinplugin.jsonmust be an object —{"type": "command", "command": "..."}— not a string. Setting it to a plain string causes Claude Code to silently ignore it.make validatedoes not runclaude plugin validate— the Makefile'svalidatetarget is local-only. The authoritative validator isclaude plugin validate .(requires the CLI).set -euo pipefailin smoke scripts exits on first grep non-match — health check scripts must not useset -e; they need to complete all checks before reporting.findinskills/returnsskills/toolkit/skill-creator/SKILL.mdwith a/Users/path — the workflow-agent-improved version has this; file it as a WARN, not FAIL, until it's cleaned.
Troubleshooting
| Symptom | Fix |
make validate fails on shellcheck | Run shellcheck hooks/<script>.sh and fix the flagged line |
make validate fails on JSON | Run jq empty .claude-plugin/plugin.json to get the parse error |
SKILL.md missing description: | The YAML block must have description: as its own line (not nested) |
plugin.json .statusLine is wrong type | Set "statusLine": {"type": "command", "command": "bash ${CLAUDE_PLUGIN_ROOT}/scripts/statusline.sh"} |
| Smoke test exits early | The script must not use set -euo pipefail; remove it if added |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

