Borrowing it
Nothing to install: this file belongs to abcwyc/pi-agent-desktop. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/abcwyc/pi-agent-desktop/main/AGENTS.mdgit clone --depth 1 https://github.com/abcwyc/pi-agent-desktopWrote 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/abcwyc/pi-agent-desktop/agents-md)<a href="https://agentmods.dev/instructions/abcwyc/pi-agent-desktop/agents-md"><img src="https://agentmods.dev/badge/instructions/abcwyc/pi-agent-desktop/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/abcwyc/pi-agent-desktop/agents-md"><img src="https://agentmods.dev/badge/instructions/abcwyc/pi-agent-desktop/agents-md.svg" alt="Reviewed on agentmods" width="80" 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 | $0.08061 | $0.08061 |
| Opus 5.5 | $0.03224 | $0.03224 |
| Sonnet 5.5 | $0.01612 | $0.01612 |
| Haiku 4.5 | $0.00806 | $0.00806 |
Grade B, and why
pi-agent-desktop AGENTS.md scanned grade B with 1 finding 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.
Reads MCP configurationmediumAgent snooping
mcp.json carries server URLs and auth tokens; reading it lets a mod discover and abuse other integrations.
- [mcp-settings.md](docs/agents/mcp-settings.md): Settings › MCP reads without running anything, masking, the trust dialog's server list, row states, notices, Code mode choice, trust from Settings, Escape stacking, every Copies of this mod
1 near-identical copy found in the catalogue:
- omp-web AGENTS.md — 88% identical, 279 lines differ
How it starts
The opening of the file, as written. The whole thing — 307 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Pi Web - Development Notes
Quick Start
npm run dev # port 30141
Typecheck: node_modules/.bin/tsc --noEmit · Lint: npm run lint
Never run next build during dev: it pollutes .next/ and breaks npm run dev.
Local environment traps
- Building (or starting a fresh
next dev) from inside the Pi desktop app's shell fails. The app exportsNODE_ENV=production,__NEXT_PRIVATE_ORIGINand__NEXT_PRIVATE_STANDALONE_CONFIG(its own serialized Next config, from which function values are gone). Next's CLI prefers that serialized config overnext.config.ts, sonext builddies withTypeError: generate is not a functionand a freshnext devwithMissing field turbopackMemoryEviction— while a dev server that was already running keeps working, which makes it look like a repo bug.scripts/prepare-desktop.mjsdeletes both__NEXT_PRIVATE_*vars for this exact reason; do the same (or run from a normal terminal) before anynext build/npm run test:e2e. - The
nodeonPATHmay not be the one the project runs on: the Pi app bundles its own runtime, while a dev server started from a terminal uses another install. Check withps -o command= -p $(lsof -nP -iTCP:30141 -sTCP:LISTEN -t). npm run test:e2erefuses to start while a dev server holds.next/dev/lock, andE2E_SERVER_MODE=start(the mode CI uses) needs a production build. Both can be satisfied without touching the dev state by building into the isolated desktop dir and pointing the harness at it:PI_WEB_DESKTOP_BUILD=1 node_modules/next/dist/bin/next build --webpackthenE2E_SERVER_MODE=start PI_WEB_DESKTOP_BUILD=1 node e2e/run.mjs(next.config.tsswapsdistDirto.next-desktop, which is gitignored).- A worktree's
node_modulescan lag the lockfile (created before an SDK bump), which fails tests that pin newer SDK behaviour.npm testrunsnpm run check:depsfirst and stops with the mismatched packages;npm cifixes it. - Windows test fixtures use
scripts/test-platform.mjs: directory links are junctions, unavailable file links skip only their own subtests, and POSIX mode checks remain active on POSIX hosts. Shell-command fixtures select native Git Bash inside the test process; trust fixtures must not inherit the user's ancestor skills. Await session shutdown before deleting MCP working directories. Real PTY integration runs in a child process; fake-clock tests use a controlled PTY. next devrewritesAGENTS.md: it appends a "This is NOT the Next.js you know" block to this file on every start. Preserve the project and user instructions here; review generated additions separately and never reset the entire file to discard development output.
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 Changed · +6 lines · +256 tokens per session 2e03b17cab79
- 4d ago Changed · +5 lines · +116 tokens per session 12cead04fe17
- 5d ago Changed · +10 lines · +188 tokens per session 90c6cfb0465e
- 6d ago Changed · +5 lines · +134 tokens per session 638bea26d3fb
- 8d ago Changed · +6 lines · +99 tokens per session 7e9e7a64c922
- 15d ago Changed · +72 tokens per session 466224784129
- 17d ago Changed · +60 tokens per session e0ee52583180
- 18d ago Changed · +12 lines · +186 tokens per session ce3cb1199166
- 19d ago Changed · +2 lines · +136 tokens per session fce69b3665ef
- 20d ago First seen · 261 lines · 6,814 tokens per session scan B 7a21ddf450c0
pi-agent-desktop AGENTS.md is an instructions file published in the GitHub repository abcwyc/pi-agent-desktop (900 stars, last pushed 2d ago), licensed MIT. It adds 8,061 tokens to every session, about $0.0322 per session on Opus 5.5. A static security scan graded it B with 1 finding (reads mcp configuration). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
pi-agent-desktop AGENTS.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| pi-agent-desktop AGENTS.mdabcwyc/pi-agent-desktop | Instructions file | 900 | 2d ago | MIT | B | 8,061 | Codex, OpenCode |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | yesterday | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | yesterday | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 274k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |