pi-council AGENTS.md

Project instructions for pi-council, a tool that starts separate agent processes to answer a question and compares their responses. They describe its main files, public tools, runtime behavior, and design rules.

In plain words
What is it for?
Use them when working on council orchestration, member processes, configuration, saved runs, follow-up questions, cancellation, status, result reading, or its MCP server.
Why use it?
They give contributors a concise map of the codebase and the rules to follow when changing or extending the project.

Instructions file for CodexOpenCode

Install

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.

agentmods
npx agentmods add instructions/sshkeda/pi-council/agents-md
Clone the repo
git clone --depth 1 https://github.com/sshkeda/pi-council

Made for: Codex, OpenCode.

Per session 610 This file is loaded in full into every session.
When invoked 610 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00610 $0.00610
Opus 5 $0.00305 $0.00305
Sonnet 5 $0.00122 $0.00122
Haiku 4.5 $0.00061 $0.00061

Measured 3d ago against content hash 55b1e67bcb43, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

pi-council 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 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.

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.

AGENTS.md · 63 lines

How it starts

The opening of the file, as written. The whole thing — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.

pi-council agent notes

pi-council spawns independent pi --mode rpc member processes and collects their answers for comparison. Keep docs concise and avoid duplicating README prose here.

Main files

  • src/core/council.ts — run lifecycle, member orchestration, persistence, follow-ups, cancellation.
  • src/core/member.ts — RPC process wrapper for one pi member.
  • src/core/config.ts~/.pi-council/config.json loading, validation, profile/model resolution.
  • src/core/runs.ts and src/core/storage.ts — persisted run lookup and cleanup.
  • src/core/types.ts — shared public types.
  • extensions/pi-council/index.ts — pi extension tools and status widget.
  • src/mcp/server.ts — MCP tool server.
  • skills/pi-council/SKILL.md — skill instructions shown to pi agents.
  • config.default.json and config.schema.json — user config template and schema.

Public tools

Pi extension:

spawn_council({ question, profile?, models?, label? })
council_followup({ message, type, runId?, memberIds? })
cancel_council({ runId?, memberIds? })
council_status({ runId? })
read_council_stream({ runId?, memberId })

MCP also exposes:

list_council_runs({})
read_council_results({ runId?, wait?, timeoutMs? })
cleanup_council_runs({ runId? | all? })

Runtime behavior

  • Config is read from ~/.pi-council/config.json; there is no config CLI.
  • Profiles select model IDs from the top-level models map and may set systemPrompt, thinking, and memberTimeoutMs.
  • Deprecated top-level systemPrompt is still used as a fallback for profiles without one.
  • Model names are passed through unchanged; use concrete model IDs from the local pi install.
  • Run artifacts are stored in ~/.pi-council/runs/<run-id>/.
  • Members are launched with PI_COUNCIL_MEMBER=1; the extension must not register council tools in that environment.

Design rules

  • Preserve model independence. Do not share one member's analysis with another unless the user explicitly sends a follow-up.
  • Prompt neutrally. The council's value is disagreement, not confirmation of the orchestrator's opinion.
  • Do not poll after spawning in interactive sessions; results are auto-delivered.
  • Persist per-member results as soon as each member finishes.
  • Avoid adding hard-coded model defaults in TypeScript; use config.default.json for templates.

Read the full file on GitHub · 63 lines

Changes

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.

  1. 3d ago First seen · 63 lines · 610 tokens per session scan A 55b1e67bcb43

Subscribe to this mod's changes

pi-council AGENTS.md is an instructions file published in the GitHub repository sshkeda/pi-council (5 stars, last pushed 4mo ago), licensed MIT. It adds 610 tokens to every session, about $0.0030 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.