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/msshives-gif/claude-plugins/agents-mdgit clone --depth 1 https://github.com/msshives-gif/claude-pluginsWrote 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/msshives-gif/claude-plugins/agents-md)<a href="https://agentmods.dev/instructions/msshives-gif/claude-plugins/agents-md"><img src="https://agentmods.dev/badge/instructions/msshives-gif/claude-plugins/agents-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 | $0.00380 | $0.00380 |
| Opus 5 | $0.00190 | $0.00190 |
| Sonnet 5 | $0.00076 | $0.00076 |
| Haiku 4.5 | $0.00038 | $0.00038 |
Grade A, and why
claude-plugins AGENTS.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 4d 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-plugins ("claude context tools")
Marketplace monorepo: every plugin lives under plugins/<name>/, one
marketplace at .claude-plugin/marketplace.json. Simple code and
simple documentation is best; complexity and big words are the enemy.
Prefer the boring obvious shape; build machinery only when real use
forces it.
- All plugins parse undocumented Claude Code internals (transcript JSONL, sessions registry, meta.json). Never make parsing stricter without keeping degrade-to-silence; when the observed format changes, update the affected plugins' fixtures TOGETHER — this shared surface is why the repo is a monorepo.
- Shared measurement core:
plugins/subagent-context/hooks/ subagent_context.py, vendored into send-guard bytools/sync-core.py(drift-guard test enforces sync). compact-manager owns its hand-ported copy outright. - Hooks MUST fail open (exit 0 on every path) and each plugin's
uninstall markers must match only its own hook paths — plugins
share the repo directory name (see each plugin's
tests/test_install_scripts.py). - Keep the repo system-agnostic: no machine-specific paths, hosts, or identities in tracked files; managed mode is Linux/WSL-only and says so rather than pretending otherwise.
- Per-plugin discipline lives in each plugin's own CLAUDE.md/README.
- Unit suites:
python3 -m unittest discover plugins/<name>/testsper plugin. compact-manager's managed mode additionally has a gated LIVE suite (tools/live-managed/run.sh— spawns a real claude session; run manually) that must pass before releasing changes to the watcher/ladder.
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.
- 4d ago First seen · 31 lines · 380 tokens per session scan A 6c3462a2a4b3
claude-plugins AGENTS.md is an instructions file published in the GitHub repository msshives-gif/claude-plugins (2 stars, last pushed 7d ago), licensed MIT. It adds 380 tokens to every session, about $0.0019 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
workframe CLAUDE.md
Instructions for ryanzhao1011/workframe, covering workframe — 仓库维护约定, 这是什么, 四条硬约束, 两份 changelog and 提交与平台.
mumei CLAUDE.md
Instructions for iroha924/mumei, covering mumei development guide, what mumei is, language policy, bash + jq conventions (essentials) and schemas/ conventions (essentials).
handoff CLAUDE.md
Instructions for AbdurRafay2004/handoff, covering claude.md, what this repo is, runtime model (how the pieces connect), invariants — do not break these when editing hooks and working / testing the hooks.
claude-sethu CLAUDE.md
Claude Code instructions for NamrataAShettar/claude-sethu, covering sethu — working notes for claude, code map, before any change, gotchas & context and release.
Slime AGENTS.md
AGENTS.md instructions for bitqs/Slime, covering agents.md — read this first, then you can ship, what slime is, the prime directive (never violate), the other hard rules and where things live.
Slime CLAUDE.md
Claude Code instructions for bitqs/Slime, covering claude.md, what this is, commands, eyeball the arena without a real session and open http://127.0.0.1:4118 (?calm=1 = flash-free).