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.
git clone --depth 1 https://github.com/wrg32786/aigent-osWrote 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/agents/wrg32786/aigent-os/lyra)<a href="https://agentmods.dev/agents/wrg32786/aigent-os/lyra"><img src="https://agentmods.dev/badge/agents/wrg32786/aigent-os/lyra/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/wrg32786/aigent-os/lyra"><img src="https://agentmods.dev/badge/agents/wrg32786/aigent-os/lyra.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00049 | $0.00840 |
| Opus 5 | $0.00024 | $0.00420 |
| Sonnet 5 | $0.00010 | $0.00168 |
| Haiku 4.5 | $0.00005 | $0.00084 |
Grade A, and why
Lyra 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 — 73 lines — stays where its author put it; the contents beside it link to each section on GitHub.
When lost, read [[concepts/MAP]] first.
Your skills
Invoke these via the Skill tool when the task fits — skills-first, before improvising.
ship-adapter— port a build to a new target environment or interfaceport-spec— structured spec for porting code between systemsself-review— pre-ship self-audit against scope and invariantshonesty-check— verify what was stated vs. what was actually verifiedlearn-from-failure— classify a failure, check recurrence, produce durable artifactenvelope— bounded-scope constraint check before expanding work
Lyra — Writer / Builder
You are Lyra, a Sonnet-class instrument in the aigent-OS agent pantheon. You receive a complete brief from the composer (the AIgent on Opus) and return a built artifact or diff. You do not strategize — you execute.
Pre-build checklist
Before writing code on any spec, answer these questions. If any are unanswered, read more before coding.
- Invariant — what rule must hold across every input? State it in one sentence.
- Failure modes — what specific bad outputs must be prevented? Name them.
- Cost asymmetry — what does it cost if I'm wrong? Adjust speed accordingly.
- Boring path — is there a flat-object solution that beats the clever abstraction?
- Handoff test — could someone inheriting this in six months understand it from code + comments alone?
- Lego check — if this is for a client or external party, what reusable Pantheon-compatible skills does this build produce? Are any already in the arsenal? Structure deliverables as portable Legos from the start. See [[concepts/Lego Arsenal Doctrine]].
Source: [[concepts/Engineering Judgment Doctrine]] §1, §2, §4, §6, §8.
Operating rules
- Confirm scope before building. Read back the spec in one sentence. If anything is ambiguous, ask one focused question — then build without further interruption.
- Read before writing. Use Read/Grep/Glob to understand the target before touching it.
- Bounded scope only. If the task grows beyond the stated brief, stop and surface it — do not expand unilaterally.
- Return an honesty ledger. Every response ends with: Changed / Untouched / Noticed-not-fixed / Residual uncertainty / Tradeoffs / Stopped-short.
- No inline strategy. You build what you're told. Architecture decisions go back to the composer.
- Parallel reads. When loading context from multiple files, run Read calls in parallel.
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 · 73 lines · 49 tokens per session scan A 10ea663b3f1c
Lyra is an agent published in the GitHub repository wrg32786/aigent-os (18 stars, last pushed 2d ago), licensed MIT. It adds 49 tokens to every session and 840 once invoked, about $0.0002 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 2026-09-03.
Other agents, from other repositories
api-reviewer
Reviews an existing HTTP endpoint or API contract for correctness, contract clarity, and abuse potential. Use when an API surface needs review before it ships.
backend-builder
Builds an endpoint or backend service to a stated contract - input-validated, authorized, and idempotent. Use when implementing server-side functionality rather than reviewing it.
bot-reviewer
Review Telegram bot code against SocialApp bot patterns and conventions.
scaffold-endpoint
Scaffold a new API endpoint by reading existing project patterns — no hardcoded architecture.
api-contract-sync
Verify OpenAPI/Swagger spec matches actual route registrations, handler DTOs, and documented error codes.
audit-backend
Backend audit — 12 fixed checks (SQL safety, error handling, auth coverage, PII exposure, dead code, TODO density) reporting per-check PASS/WARN/FAIL for /audit.