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/patforna/auto-task/claude-mdgit clone --depth 1 https://github.com/patforna/auto-taskWhat 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 | $0.00437 | $0.00437 |
| Opus 5 | $0.00218 | $0.00218 |
| Sonnet 5 | $0.00087 | $0.00087 |
| Haiku 4.5 | $0.00044 | $0.00044 |
Grade A, and why
auto-task CLAUDE.md scanned grade A with 0 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 3d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
CLAUDE.md — Auto-Task Plugin
Vendored Skills — Do Not Edit Here
skills/panel, skills/synthesize, and skills/tdd are vendored copies from the core-skills plugin (a sibling checkout; scripts/vendor.sh takes its path as an argument), which is their source of truth. Do not edit them in this repo — edits here will be overwritten on the next re-vendor and silently diverge from core-skills.
To change one of them: edit it in core-skills, then re-vendor here by running scripts/vendor.sh. It copies the three dirs and rewrites /core-skills: → /at: — the only way the vendored copies differ from the originals (keeping this plugin self-contained). Then bump the version in .claude-plugin/plugin.json + .claude-plugin/marketplace.json and refresh the installed plugin: claude plugin marketplace update auto-task && claude plugin update at@auto-task.
Internal Skill References
This plugin is self-contained: its skills invoke each other via the /at:<skill> namespace (never bare /<skill>), so they resolve regardless of what else is installed. The only external reference is the codex plugin (/codex:*), which is optional — skills must degrade gracefully (documented fallback) when it is not installed.
Portability
Skills must stay project-agnostic: generic instructions with sensible defaults, and any project-specific behaviour sourced from the consuming repo's config — .claude/auto-task.config.md (project) and .claude/auto-task.config.local.md (personal overrides, win on conflict). Never hardcode a specific project's paths, recipes, conventions, or context into a skill body.
When adding or renaming a config setting, update all three places in the same change: the skill(s) that consume it, examples/auto-task.config.md (the copy-paste template), and the README's config table.
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.
- 3d ago First seen · 18 lines · 437 tokens per session scan A fae11f963c39
auto-task CLAUDE.md is an instructions file published in the GitHub repository patforna/auto-task (3 stars, last pushed 14d ago), licensed MIT. It adds 437 tokens to every session, about $0.0022 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
superpowers GEMINI.md
Gemini CLI instructions for obra/superpowers, a project described as: An agentic skills framework & software development methodology that works.
turbo-flow CLAUDE.md
Claude Code instructions for marcuspat/turbo-flow, covering claude.md — turbo flow, what this repo is, key files and conventions.
cdviz CLAUDE.md
Instructions for cdviz-dev/cdviz, covering claude.md, agents.md - cdviz monorepo, project context, monorepo structure and task execution pattern.
aiworkspace AGENTS.md
Instructions for a-tokyo/aiworkspace, covering ai workspace, structure, key commands, conventions and code quality.
klimkit AGENTS.md
Instructions for klimentij/klimkit, covering agents.md, work log (docs/work), html reports and explainers — hosted, auth-protected and main release reminder.
engineering-skills CLAUDE.md
Instructions for int2t05/engineering-skills, covering claude.md, conventions and skill discovery.