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/dkships/pm-copilot/claude-mdgit clone --depth 1 https://github.com/dkships/pm-copilotWrote 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/dkships/pm-copilot/claude-md)<a href="https://agentmods.dev/instructions/dkships/pm-copilot/claude-md"><img src="https://agentmods.dev/badge/instructions/dkships/pm-copilot/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 | $0.01469 | $0.01469 |
| Opus 5 | $0.00734 | $0.00734 |
| Sonnet 5 | $0.00294 | $0.00294 |
| Haiku 4.5 | $0.00147 | $0.00147 |
Grade A, and why
pm-copilot 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 5d 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.
How it starts
The opening of the file, as written. The whole thing — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PM Copilot — MCP Server
MCP server connecting an LLM to customer signal data (HelpScout support tickets + ProductLift feature requests). Architecture and quickstart: README.md. Contributor process: CONTRIBUTING.md.
Commands
npm run build # Compile TypeScript (also chmods dist/index.js)
npm run dev # Watch mode
npm test # Vitest unit tests
npm start # Run the server
npm run tool -- <tool_name> '<json-params>' # Call one tool directly, no MCP client (use --list to enumerate)
npm run eval # Theme-matching precision/recall (--failures, --json, --fixture, --min-f1)
npm run audit:ci # Dependency audit gate (prepublishOnly runs build + test + this)
Run npm run build after source changes before testing through an MCP client — clients execute dist/index.js. Restart Claude Code to pick up server changes. The server is registered via the project-scoped .mcp.json; re-add with claude mcp add pm-copilot --scope project -- node dist/index.js.
PII guardrails (load-bearing for any PR)
- Scrub patterns live in
src/pii-scrubber.ts(SSN, Luhn-validated credit cards, email addresses, phone numbers); they are applied insrc/format.tsbefore text enters analysis or leaves the server. Customer email field is always[REDACTED]regardless of match. - Never send unscrubbed customer text. New customer data sources must route through the format layer.
- Excluded by design (don't add without scoping a privacy review): agent/admin responses, internal HelpScout notes, attachments, voter identities, commenter names and emails, raw thread bodies.
kpi_contextis passed through verbatim — caller's responsibility. - Chatbase is the widest PII surface (unbounded chat free text).
formatDeflectedConversationreadsrole: "user"turns only and deliberately drops assistant turns,form_submission(captured lead PII),userIdandcountry.src/deflection.test.tsasserts none of those reach the formatted output — keep that test passing. - Every tool response includes
pii_scrubbing_applied: trueandpii_categories_redacted: [...].preview_only: trueongenerate_product_planreturns a manifest without fetching data.
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.
- 5d ago First seen · 87 lines · 1,469 tokens per session scan A 02b6355ebba3
pm-copilot CLAUDE.md is an instructions file published in the GitHub repository dkships/pm-copilot (29 stars, last pushed 11d ago), licensed MIT. It adds 1,469 tokens to every session, about $0.0073 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 2026-08-30.
Other instructions, from other repositories
mcp-jira-confluence CLAUDE.md
Instructions for thamaraiselvam/mcp-jira-confluence, covering claude.md, commands, architecture, request flow and transport modes (index.ts).
Noema AGENTS.md
Instructions for Fail-Safe/Noema, covering repository guidelines, project structure & module organization, build, test, and development commands, coding style & naming conventions and testing guidelines.
video-research-mcp AGENTS.md
Instructions for Galbaz1/video-research-mcp, covering agents.md, scope, what this is, commands and code review trigger protocol.
mcp-redmine-rest CLAUDE.md
Claude Code instructions for alsimoes/mcp-redmine-rest, covering claude.md, what this is, layout (src-layout), tool registration pattern and deliberately missing tools.
datadog-mcp CLAUDE.md
Claude Code instructions for dreamiurg/datadog-mcp, covering agent instructions, project, stack, commands and releases.
festival GEMINI.md
Gemini CLI instructions for Obedience-Corp/festival: Festival is a goal-oriented project management methodology for human and AI development workflows, driven by the fest and camp CLIs, installed and kept in sync by a third tool, festival. The skills imported below describe how to plan and execute festivals. Load them…