Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/pramoddutta/qaskills/claude-mdgit clone --depth 1 https://github.com/PramodDutta/qaskillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/pramoddutta/qaskills/claude-md)<a href="https://agentmods.dev/instructions/pramoddutta/qaskills/claude-md"><img src="https://agentmods.dev/badge/instructions/pramoddutta/qaskills/claude-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.05287 | $0.05287 |
| Opus 5 | $0.02644 | $0.02644 |
| Sonnet 5 | $0.01057 | $0.01057 |
| Haiku 4.5 | $0.00529 | $0.00529 |
Grade C, and why
qaskills CLAUDE.md scanned grade C with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
- Deleting files you did not create this session; `rm -rf` anywhere in the repo; force push; history rewrites Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
4. **The stale .env.local seed.** Seeding "production" using `.env.local`'s `DATABASE_URL`, which points at an old non-prod database; the live site never changes and new skills 404. Rule: `.env.local` is NOT production. How it starts
The opening of the file, as written. The whole thing — 222 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Operating manual for AI agents working in this repository. It is written so a model with less judgment than the one that wrote it can still ship at the same level: where a rule exists, follow the rule, do not improvise. Every named mistake below is one that actually happened or nearly happened here.
What this is
QASkills.sh is a QA-skills directory for AI coding agents ("npm for QA skills"). pnpm monorepo (pnpm 9.15.0, Node >= 20, Turborepo). Solo-founder operation: work lands directly on main, ships to production the same day, and growth is SEO-driven. The three recurring jobs, by frequency:
- Publish SEO articles (near-daily batches of 10) -> use project skill
publish-seo-batch - Add seed skills to the catalog -> use project skill
add-seed-skills - Deploy to production on Vercel -> use project skill
ship-prod
Project skills live in .claude/skills/. If a task matches one, follow the skill, not memory.
Conventions (non-negotiable)
Writing (everything: chat, commits, docs, article content, code comments):
- Never use em dashes. Use a comma, period, colon, parentheses, or "->". En dashes for ranges (Jul 14-19) are fine. Before finishing any writing deliverable, run a literal grep for the em dash character and fix hits.
Git:
- Conventional commits, imperative mood:
feat:,fix:,docs(scope):,chore:. Match the style ingit log. - No
Co-Authored-By:trailers. No "Generated with Claude Code" footers. Commits must look like normal human commits. - Commit directly to
main(repo norm, no PR ceremony). One concern per commit. - Never
git add -A,git add ., orgit commit -afrom the repo root. The root is littered with unrelated artifacts (screenshots,evaldog-app/,qabuddy/, one-off strategy docs). Stage explicit paths only, and reviewgit diff --cached --statbefore committing. - Pre-existing working-tree changes and untracked files are someone's WIP. Never stash, clean, revert, or absorb them into your commit.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago First seen · 222 lines · 5,287 tokens per session scan C c78183fa5052
qaskills CLAUDE.md is an instructions file published in the GitHub repository PramodDutta/qaskills (217 stars, last pushed 6d ago), licensed MIT. It adds 5,287 tokens to every session, about $0.0264 per session on Opus 5. A static security scan graded it C with 2 findings (recursive force delete, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
true-skills AGENTS.md
Instructions for katalon-labs/true-skills, covering agents.md - katalon true platform skills, find your skill by role, skills (single source of truth: skills/), platform-setup and test-plan.
true-skills copilot-instructions.md
Instructions for katalon-labs/true-skills, covering katalon true platform - copilot instructions, skills and katalon mcp server.
rrweb-stack AGENTS.md
Instructions for Cubenest/rrweb-stack, covering project context for ai coding assistants, what the repo is, build, test & dev commands, planning context — the context/ convention and maintainer-only setup (do this once per fresh checkout).
rrweb-stack CLAUDE.md
Instructions for Cubenest/rrweb-stack, covering claude code and deploying the docs sites.
vibe-testing CLAUDE.md
Claude Code instructions for AishwaryShrivastav/vibe-testing, covering vibe test — code-aware browser testing agent, what this project does, architecture, mcp tools (13) and development.
llm-safe-haven CLAUDE.md
Claude Code instructions for pleasedodisturb/llm-safe-haven, covering llm safe haven, what this is, project structure, tdd — non-negotiable (adopted 2026-08-17) and the contract.