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 skills/sliday/harn/agents-mdnpx skills add sliday/harn --skill agents-mdgit clone --depth 1 https://github.com/sliday/harnWhat 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 | $0.00039 | $0.00476 |
| Opus 5 | $0.00019 | $0.00238 |
| Sonnet 5 | $0.00008 | $0.00095 |
| Haiku 4.5 | $0.00004 | $0.00048 |
Grade A, and why
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 2d 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
Reference: https://harn.app/kb/specs.html — "12 Factor Agents"
Reference: https://harn.app/kb/context.html — "Writing a good CLAUDE.md"
AGENTS.md Builder
Generate a lean, effective AGENTS.md using progressive disclosure principles.
Rules
- Under 60 lines — ideally under 100, never over 300
- No directory trees — agents discover structure on their own
- No bloated context — point to skill files for domain-specific rules
- Progressive disclosure — load detailed instructions only when relevant
Process
- Read the current project root (package.json, README, existing AGENTS.md)
- Identify: stack, entry points, architectural constraints
- Generate AGENTS.md with sections:
- North Star (1-2 sentences)
- System of Record (stack, entry points)
- Constraints (hard rules)
- Active Harness Hooks (what's automated)
- Skills (progressive disclosure pointers)
- Escape Hatch (what to do when stuck)
- If existing AGENTS.md is over 100 lines, refactor it:
- Extract domain rules into separate skill files
- Keep only pointers in AGENTS.md
Workflow
- Understand → Read code, check LEARNED.md for gotchas
- Plan → Break task into steps, update CHECKPOINT.json
- Implement → Write code within architectural constraints
- Verify → Run quality gate before finishing
- Document → Update LEARNED.md if something was tricky
Context Budget
- Keep this file under 60 lines — load skills/ on demand
- Delegate research to sub-agents — only summaries return
- After 30+ tool calls, compact and update CHECKPOINT.json
Escape Hatches
- Quality gate stuck? Stop hook checks
stop_hook_active— retry lets you through - Security guard wrong? Report false positive, use alternative command
- Same error 3 times? Stop and ask the human
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.
- 2d ago First seen · 51 lines · 39 tokens per session scan A 52d5d320b14f
agents-md is a skill published in the GitHub repository sliday/harn (5 stars, last pushed 4mo ago), licensed MIT. It adds 39 tokens to every session and 476 once invoked, about $0.0002 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 skills, from other repositories
mcporter
List, auth, and call MCP servers/tools from the terminal.
📝 任务完成后归档
重要提醒: 每次完成复杂调试或开发任务后，主动执行此流程！ 将学到的经验归档为 skill，供以后参考。不要等用户提醒。.
oracle
Best practices for using the oracle CLI (prompt + file bundling, engines, sessions, and file attachment patterns).
agent-mode
Unified tool for managing agent LLM modes (add, remove, update, list, switch).
agento11y-prod-setup
Sets up production evaluation and guardrails for a DEPLOYED AI agent in Grafana Agent Observability, grounded in the agent's own code and its real ingested traffic. The judgment layer on top of the agento11y skill: it reads the agent's source (system prompt, tools, entrypoint) AND samples its live traffic via gcx…
mcp-scripting
Write mcpScript JavaScript for discovering, inspecting, and calling MCP tools.