pi-agent-dashboard AGENTS.md

pi-agent-dashboard AGENTS.md is an instructions file for Codex, OpenCode from BlackBeltTechnology/pi-agent-dashboard. It costs 3,679 tokens per session, scanned A, original, MIT.

Project instructions for PI Dashboard, a web dashboard that monitors and controls pi AI-agent sessions remotely. They define the system architecture and require agents to consult the project's knowledge tools before searching code or reading files.

In plain words
What is it for?
Working on the dashboard's bridge, server, and React web client; looking up file purposes and exports; tracing documented relationships; and following the project's investigation and documentation-update rules.
Why use it?
They enforce a documentation-first investigation process, so agents learn a file or feature's purpose before tracing implementation details. This reduces blind searches and helps keep changes aligned with the documented design.

Instructions file for CodexOpenCode

Written for Codex and OpenCode: the file is AGENTS.md. Also seen: mentions subagents; mentions AGENTS.md.

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/blackbelttechnology/pi-agent-dashboard/agents-md
Clone the repo
git clone --depth 1 https://github.com/BlackBeltTechnology/pi-agent-dashboard

Made for: Codex, OpenCode.

Wrote 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.

agentmods badge for pi-agent-dashboard AGENTS.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/blackbelttechnology/pi-agent-dashboard/agents-md.svg)](https://agentmods.dev/instructions/blackbelttechnology/pi-agent-dashboard/agents-md)
Your own site
<a href="https://agentmods.dev/instructions/blackbelttechnology/pi-agent-dashboard/agents-md"><img src="https://agentmods.dev/badge/instructions/blackbelttechnology/pi-agent-dashboard/agents-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 3,679 This file is loaded in full into every session.
When invoked 3,679 The same file — it is already loaded in full.
Security scan A 1 finding. 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 5.1 $0.03679 $0.03679
Opus 5 $0.01840 $0.01840
Sonnet 5 $0.00736 $0.00736
Haiku 4.5 $0.00368 $0.00368

Measured yesterday against content hash e46f3425a1e1, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

pi-agent-dashboard AGENTS.md scanned grade A 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 yesterday.

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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -X POST http://localhost:8000/api/restart # after packages/server|packages/shared changes (jiti — no build)
AGENTS.md · 146 lines

How it starts

The opening of the file, as written. The whole thing — 146 lines — stays where its author put it; the contents beside it link to each section on GitHub.

PI Dashboard

Web dashboard to monitor + control pi agent sessions remotely. Three components: bridge extension + Node server + React client. Full architecture: docs/architecture.md.

Docs-First Gate — kb before grep (per-turn doctrine)

kb_* tools return a one-line purpose + key exports per file instead of raw bytes. This gate fires on the ACTION, not the intent — before you grep/rg a symbol, cat/Read a file to learn its purpose, or chase an import, the kb call goes first. It fires even mid-task when you already know the file. When your reflex is the left column, run the right instead:

You're about to… Do this FIRST instead
grep -rn "SymbolName" — find where a fn/type/const lives kb_search --doc-type agents "SymbolName"
grep -rn "topic" src/ — how does X work / where's X handled kb_search "feature topic"
cat/Read a file to learn its purpose before editing kb agents <path> — purpose + exports + See change:
chase imports / callers across files rg "<symbol>" — the Tier-1 graph is markdown-structure only and CANNOT resolve code refs
read one doc section in full kb_get <path> <section>
a kb hit shows STALE / GONE / UNVERIFIED (trust verdict) verify the row against source before acting — MOVED verifies at its reported successor path; FRESH may be acted on without re-reading; see kb_search verdicts (trust label, never ranking)
build / run / install / setup / release / "how do I X" grep -i <kw> docs/faq.md README.md docs/ — then quote
derive a fact from a large file / big command output ctx_execute_file / ctx_execute when present (context-mode is optional); else Read w/ offset+limit, or rg/awk via Bash

kb_search indexes repo markdown (docs/ openspec/ packages/ .pi/) — NOT tests/ qa/ scripts/ docker/. ctx_search/memory_search index session memory, NOT repo docs — different corpus.

Pick the lane — this is the single highest-yield kb habit. Looking for a FILE or SYMBOL → pass doc_type:"agents" (measured P@1 0.048 → 0.231, MRR 0.187 → 0.327 on the 104 mined file-lookup queries; unfiltered, verbose openspec/ spec prose takes rank 1 and buries the per-file row at rank 5-10). Asking how something WORKS, or anything conceptual → leave doc_type unset; the agents filter measurably HURTS prose queries (P@1 0.151 → 0.068, R@10 0.575 → 0.205). Reproduce: tsx packages/kb/eval/run-fixtures.ts.

Read the full file on GitHub · 146 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. yesterday Changed · +1 lines · +39 tokens per session e46f3425a1e1
  2. 2d ago Changed · +23 tokens per session 72daa9044e2b
  3. 6d ago First seen · 145 lines · 3,617 tokens per session scan A a9b26055cc26

Subscribe to this mod's changes

pi-agent-dashboard AGENTS.md is an instructions file published in the GitHub repository BlackBeltTechnology/pi-agent-dashboard (274 stars, last pushed today), licensed MIT. It adds 3,679 tokens to every session, about $0.0184 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.