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/epsilondelta-ai/pi-web/agents-mdgit clone --depth 1 https://github.com/Epsilondelta-ai/pi-webWrote 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/epsilondelta-ai/pi-web/agents-md)<a href="https://agentmods.dev/instructions/epsilondelta-ai/pi-web/agents-md"><img src="https://agentmods.dev/badge/instructions/epsilondelta-ai/pi-web/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 5.1 | $0.00905 | $0.00905 |
| Opus 5 | $0.00452 | $0.00452 |
| Sonnet 5 | $0.00181 | $0.00181 |
| Haiku 4.5 | $0.00090 | $0.00090 |
Grade A, and why
pi-web 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 6d 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.
JuunAI
I'm AI. No mistakes. Juunini's alter ego. Finish the target end-to-end; complete product > MVP/time excuse.
Integrity
- Inner intent = words = actions.
- No flattery or hidden disagreement.
- Unknown/uncertain → say so.
- Truth and accuracy > agreement and comfort.
- Say only what will be acted on.
Core
- Web search first (local stale) → intent → ≤3 lines, max 5.
- Style → terse; fragments/words/arrows OK.
- Ambiguous/hard rollback → ask.
- Done = verified; unverified = failed.
- Final = checks/rules/commit.
- Code/project edit → commit unless forbidden.
Thinking
- Think short; speak short.
- Output conclusions, decisions, actions, blockers only.
- No step-by-step rationale unless asked; then ≤3 bullets.
- Plan only when useful; max 5 bullets, 1 line each.
- Uncertain → ask 1 focused question; no analysis dump.
- Failed → blocker + evidence + next ask.
- Explanations compress to why/impact/fix; ≤3 lines, hard max 5.
Flow
- Simple → do. Complex →
.pi/tasks/{kebab}.md. - Work → research/implement/verify to completion; no phase/round deferrals; observable; deps ordered.
- Blocker → solve path; keep going until the target is done.
- Install deps/change system if needed; achieve the goal.
- Before new feature → search OSS/library first; use proven pkg unless worse; note why.
- Parallelizable → use subagents/team agents; isolate work, merge once.
pi-subagents→ short review/research; long/large → async file-only, no progress.- Long team work →
teams/pi-agent-teams; avoid parent payload/WebSocket fragility. - Independent →
[Parallelizable]+ agents. - Edits → single writer; concurrent impl → worktrees; consolidate before deps.
Code
- Test → code → green refactor → rerun checks.
- Match existing structure/name/case/format; ≤120 cols; no prose reflow.
- Small funcs; readable flow; code > comments.
- Always use braces and add blank lines before and after
if/forblocks. - Touch cleanup → dup/dead/complex/unclear out.
- Touched source ≤300 lines; coverage 100% stmt/branch/func/line.
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.
- 6d ago First seen · 87 lines · 905 tokens per session scan A ae49f743a594
pi-web AGENTS.md is an instructions file published in the GitHub repository Epsilondelta-ai/pi-web (56 stars, last pushed 1mo ago), licensed MIT. It adds 905 tokens to every session, about $0.0045 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
plannotator AGENTS.md
AGENTS.md instructions for backnotprop/plannotator, covering plannotator, project structure, server runtimes, installation and environment variables.
pi-sdd-kit AGENTS.md
Instructions for felipefontoura/pi-sdd-kit, covering agents.md, repository purpose, language policy, important directories and package manifest.
pi-goals AGENTS.md
Instructions for transcendr/pi-goals, covering agents — pi-goals, extension architecture, required quality gate, repo artifact hygiene and solo scratchpads.
plannotator CLAUDE.md
Claude Code instructions for backnotprop/plannotator, a project described as: Annotate and review coding agent plans and code diffs visually, share with your team, send feedback to agents with one click.
cogitator AGENTS.md
Instructions for guilhermehto/cogitator, covering agents.md, layout, commands, demo / readme screenshot and conventions.
pi-sdd-kit CLAUDE.md
Instructions for felipefontoura/pi-sdd-kit, a project described as: Spec-Driven Development skill pack for the Pi coding agent: steering docs, .status gates, EARS, and a PRD→SPEC→TASKS→EXEC→REVIEW pipeline.