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/broomva/symphony/agents-mdgit clone --depth 1 https://github.com/broomva/symphonyWhat 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 | $0.01822 | $0.01822 |
| Opus 5 | $0.00911 | $0.00911 |
| Sonnet 5 | $0.00364 | $0.00364 |
| Haiku 4.5 | $0.00182 | $0.00182 |
Grade A, and why
symphony 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 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.
How it starts
The opening of the file, as written. The whole thing — 157 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md - Symphony
Repository Purpose
Symphony is a long-running orchestration service that polls an issue tracker (Linear), creates isolated per-issue workspaces, and runs coding agent sessions (Claude, Codex, etc.) for each issue. It is a scheduler/runner, not a workflow engine.
Architecture
Rust workspace with layered crates matching the spec's abstraction levels:
| Crate | Spec Layer | Responsibility |
|---|---|---|
symphony-core |
Domain Model (S4) | Shared types: Issue, State, Session, Workspace |
symphony-config |
Config + Policy (S5-6) | WORKFLOW.md loader, typed config, file watcher |
symphony-tracker |
Integration (S11) | Linear GraphQL client, issue normalization |
symphony-workspace |
Execution (S9) | Per-issue directory lifecycle, hooks, safety invariants |
symphony-agent |
Execution (S10) | Agent subprocess, JSON-RPC + simple pipe modes |
symphony-orchestrator |
Coordination (S7-8) | Poll loop, dispatch, reconciliation, retry, drain |
symphony-observability |
Observability (S13) | Structured logging, HTTP server, dashboard, health, auth |
symphony-cli (root) |
CLI (S17.7) | Subcommands: start, stop, status, issues, validate, etc. |
Key Design Decisions
- In-memory state: Orchestrator state is intentionally in-memory; recovery is tracker-driven
- Single authority: Only the orchestrator mutates scheduling state
- Workspace isolation: Coding agents run ONLY inside per-issue workspace directories
- Dynamic reload: WORKFLOW.md changes are detected and re-applied without restart
- Liquid-compatible templates: Strict variable/filter checking for prompt rendering
- Graceful shutdown: SIGTERM/SIGINT → drain mode → wait for workers → exit
- Stall kill: Worker abort handles tracked; stalled sessions killed + retried
- Bearer auth: Optional
SYMPHONY_API_TOKENprotects/api/v1/*; health endpoints open
Gathering Context from the Knowledge Graph
This repo is an Obsidian vault. All .md files form a wikilinked knowledge graph.
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.
- 2d ago First seen · 157 lines · 1,822 tokens per session scan A da1f7f21d81e
symphony AGENTS.md is an instructions file published in the GitHub repository broomva/symphony (9 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 1,822 tokens to every session, about $0.0091 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.
Other instructions, from other repositories
autoany CLAUDE.md
Instructions for broomva/autoany, covering autoany — egri framework, project structure, commands, adapter crates (standalone, not in workspace) and architecture rules.
sample-autonomous-cloud-coding-agents AGENTS.md
AGENTS.md instructions for aws-samples/sample-autonomous-cloud-coding-agents, covering agents.md, your role, commands (run these), git workflow and where to make changes.
sample-autonomous-cloud-coding-agents CLAUDE.md
Claude Code instructions for aws-samples/sample-autonomous-cloud-coding-agents, a project described as: Autonomous background coding agents on AWS. Turn tasks into pull requests via isolated runtimes, with built-in orchestration, observability, and governance.
enterprise-software-playbook AGENTS.md
Instructions for bricerising/enterprise-software-playbook, covering agent instructions (enterprise-software-playbook), repo structure, start here, change rules (high signal) and verification.
archcodex CLAUDE.md
Instructions for ArchCodexOrg/archcodex, covering claude code instructions for archcodex, before modifying code, workflow, mcp tools and understanding context output.
OwnPilot AGENTS.md
Instructions for ownpilot/OwnPilot, covering ownpilot, architecture, key patterns, commands and tech stack.