copilotd copilot-instructions.md

Project instructions for copilotd, a .NET background program that repeatedly checks GitHub issues and running processes, then starts, retries, or stops matching work sessions.

In plain words
What is it for?
Use them when building or running copilotd, investigating session management, or manually checking its status and reconciliation loop.
Why use it?
They explain how the program keeps its saved state, GitHub results, and operating-system processes aligned, including how to build and verify changes.

Instructions file for GitHub Copilot

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/damianedwards/copilotd/copilot-instructions
Clone the repo
git clone --depth 1 https://github.com/DamianEdwards/copilotd

Made for: GitHub Copilot.

Per session 1,292 This file is loaded in full into every session.
When invoked 1,292 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 $0.01292 $0.01292
Opus 5 $0.00646 $0.00646
Sonnet 5 $0.00258 $0.00258
Haiku 4.5 $0.00129 $0.00129

Measured 2d ago against content hash 65be377aa9ca, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

copilotd copilot-instructions.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 2d 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.

.github/copilot-instructions.md · 98 lines

How it starts

The opening of the file, as written. The whole thing — 98 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Copilot Instructions for copilotd

Build & Run

# Build
dotnet build src/copilotd/copilotd.csproj

# Run from source (auto-builds and passes all args through)
./copilotd.sh <command>          # macOS/Linux
.\copilotd.ps1 <command>           # Windows (PowerShell)
.\copilotd.cmd <command>           # Windows (CMD)

No test suite exists yet. Verify changes manually via copilotd status, copilotd session list, and copilotd run --interval 15 --log-level debug.

Architecture

Reconciliation-loop daemon. The core is ReconciliationEngine.Reconcile() which runs every poll cycle with five steps:

  1. Prune terminal sessions older than 7 days
  2. Verify tracked processes are alive (PID + start time to detect PID reuse)
  3. Query GitHub for issues matching rules via gh CLI
  4. Reconcile desired vs. observed state (terminate unmatched, create/retry matched)
  5. Dispatch pending sessions up to MaxInstances limit

State converges through this loop — every inconsistency (crashed process, stale session, failed API call) self-heals on the next cycle.

Three truth sources are reconciled each cycle: persisted state (~/.copilotd/state.json) → live OS process status → current GitHub issue matches.

Session lifecycle states: Pending → Dispatching → Running → Completed/Failed/Orphaned, plus WaitingForFeedback (session paused, waiting for new issue comments) and WaitingForReview. Joined is a legacy compatibility state from older versions that used local interactive takeover. CompletedBySession flag distinguishes explicit copilot completion from automatic (issue unmatched) completion.

Key Conventions

PowerShell Safety

PowerShell variable names are case-insensitive, so $home resolves to the built-in read-only $HOME variable. Never use $home as a scratch or target variable name in scripts. Use specific names relative to the context at hand instead, especially before destructive operations like Remove-Item -Recurse.

System.CommandLine v2.0.5 API

Read the full file on GitHub · 98 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. 2d ago First seen · 98 lines · 1,292 tokens per session scan A 65be377aa9ca

Subscribe to this mod's changes

copilotd copilot-instructions.md is an instructions file published in the GitHub repository DamianEdwards/copilotd (59 stars, last pushed 5d ago), licensed MIT. It adds 1,292 tokens to every session, about $0.0065 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.