hashintel/labs/harden
- community
- ext:[{"type":"hook","description":"Pre/post lifecycle hooks"}]
Checking that the SKILL.md file is still in the repository.
Deliberately upgrade the project's feedback harness. Use when adding a verify command, setting up CI, introducing lint/format, or making the dev loop self-enforcing. Covers canonical verify scripts, lint/format, CI configuration, and dev-loop documentation.
Install
npx vskill@latest install hashintel/labs/hardenSemi-extensibleMentions customization, but needs skill-specific knowledge. Read the extensible skills standard
SKILL.md
Harden
Deliberately upgrade the project's feedback harness — the commands, gates, and automation that make verification self-enforcing. Introduce lint, formatting, CI, and a canonical verify command intentionally and incrementally.
This resolves the tension between /pragma:skeleton ("no lint/format yet") and /pragma:slice ("run the verification harness").
Input
What to harden: $ARGUMENTS
Preconditions:
- A walking skeleton must exist. If none →
/pragma:skeletonfirst. - Ideally 1-2 real slices have landed (you want to know what checks actually matter before formalizing).
Procedure
1. Audit the current harness
What commands exist? Are they consistent? Is there a single "everything is green" command? Is there CI? Does CI match local?
2. Define the target verify command
One canonical command to trust:
Runs in order: typecheck → lint → format check → unit tests → integration tests → build.
- Must be fast enough for pre-commit (target: < 60s)
- If some checks are slow, split:
verify(fast) andverify:full(CI-only)
3. Introduce ONE tool
Pick one improvement per /pragma:harden invocation:
- Formatting: choose one formatter, configure minimally (defaults first), run on whole codebase, add to verify
- Lint: choose one linter, default ruleset, fix existing violations, add to verify
- CI: create config running the verify command, fail on verify failure
- Pre-commit hook (optional): must be fast (< 10s) or it will be bypassed
4. Document the dev loop
Create or update a short section (README or docs/dev-loop.md):
- How to verify (canonical command)
- What it checks (ordered list)
- How to run subsets
- CI: where it runs, what triggers it
5. Verify and commit
Run the new verify command. If the new tool creates violations, fix them in this commit. Commit: Harden: add [tool/command] to verification harness
Constraints
- One tool at a time. Do not batch formatting + lint + CI in one invocation.
- Local and CI must agree. No "CI-only" checks developers can't run locally.
- Fast by default. Slow checks go in
verify:full. - No speculative rules. Add rules that catch real problems you've seen.
- Minimal configuration. Defaults first. Customize only when a default causes real pain.
- Don't block feature work. Too much friction → back it out and try again later.
Output
- Tool added: what was introduced and why
- Verify command: current canonical command and what it runs
- Violations fixed: what existing code needed cleanup
- Dev loop doc: where the documentation lives
- Recommended next harden step: what to add next (if anything)
- Lifecycle:
State: governance,Next: /pragma:consult,Loop: /pragma:consult(default unless user explicitly continues directly)
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingpocs/agent-plugins/claude-pragma/skills/harden/SKILL.md
Popularity
- Repository stars
- 35
- Repository forks
- 14
- 7-day trend score
- 18
Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 2 Jun 2026, 4 months ago.
Extension points
hookPre/post lifecycle hooks
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/hashintel/labs/harden)