Borrowing it
Nothing to install: this file belongs to tcconnally/perseus-amd-act-ii. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/tcconnally/perseus-amd-act-ii/main/AGENTS.mdgit clone --depth 1 https://github.com/tcconnally/perseus-amd-act-iiWrote 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/instructions/tcconnally/perseus-amd-act-ii/agents-md)<a href="https://agentmods.dev/instructions/tcconnally/perseus-amd-act-ii/agents-md"><img src="https://agentmods.dev/badge/instructions/tcconnally/perseus-amd-act-ii/agents-md/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/instructions/tcconnally/perseus-amd-act-ii/agents-md"><img src="https://agentmods.dev/badge/instructions/tcconnally/perseus-amd-act-ii/agents-md.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 | $0.00769 | $0.00769 |
| Opus 5.5 | $0.00308 | $0.00308 |
| Sonnet 5.5 | $0.00154 | $0.00154 |
| Haiku 4.5 | $0.00077 | $0.00077 |
Grade A, and why
perseus-amd-act-ii 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 2mo 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 — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for AI coding agents (and humans) working in this repository, plus a description of the agent this project actually ships.
The agent this repo ships
src/agent_memory_demo.py is a minimal but complete stateful agent:
- Inference is an open-weight model behind the Fireworks AI API (target deployment: AMD Instinct via ROCm/vLLM; no serving API attests which accelerator handles a request). Inference is stateless — the context window dies with the session.
- Memory is provided by Perseus Vault, an MCP-native, local-first,
encrypted memory engine (single Rust binary). The agent
remembers durable facts, then in a later sessionrecalls them before prompting the model, so knowledge survives across sessions. - Lifecycle: a
decaytick ages rarely-used memories and archives noise (never deletes — the journal stays auditable), so recall quality holds as the store grows.
The key architectural claim: memory lives on the host CPU, not in GPU HBM. See docs/ARCHITECTURE.md.
Repo map
| Path | What it is |
|---|---|
src/perseus_vault_store.py |
The memory interface. ReferenceStore (SQLite/FTS5, CPU, always runs) and BinaryStore (bridge to a real perseus-vault binary when PERSEUS_VAULT_BIN is set). |
src/agent_memory_demo.py |
End-to-end agent: learn → new session → recall → infer → load → decay. |
src/gemma_on_amd.py |
Bonus: same loop with Gemma 3 served locally by llama.cpp on an AMD CPU (partner challenge). |
src/benchmark.py |
Measured throughput + footprint tables (1K/10K/100K) and published-spec economics. |
src/economics.py |
The "one MI300X serves N agents" math. Every value tagged published-spec / projection / measured. |
docs/ |
Architecture, benchmarks, and the pre-filled lablab submission. |
Ground rules for agents editing this repo
- The honesty rule is load-bearing. Never present a GPU/MI300X number as
measured. Every benchmark row carries a
data_sourceofmeasured,published-spec, orprojection. If you add a number, tag it and cite it. - No secrets in code. Credentials come from
.env(gitignored) via.env.example. Never inline an API key or token. - Keep the core stdlib-only. The demo and benchmark must run with no
third-party install so any judge can reproduce them. Optional deps go in
requirements.txtbehind clear comments. - Everything must stay runnable offline. No network call may be required for
the demo to complete; the Fireworks path is opt-in via
FIREWORKS_API_KEY. - MIT-licensed and original. This is a hackathon submission requirement.
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.
- 2mo ago First seen · 58 lines · 769 tokens per session scan A d705b2e35288
perseus-amd-act-ii AGENTS.md is an instructions file published in the GitHub repository tcconnally/perseus-amd-act-ii (1 stars, last pushed 3mo ago), licensed MIT. It adds 769 tokens to every session, about $0.0031 per session on Opus 5.5. 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.
perseus-amd-act-ii AGENTS.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| perseus-amd-act-ii AGENTS.mdtcconnally/perseus-amd-act-ii | Instructions file | 1 | 3mo ago | MIT | A | 769 | Codex, OpenCode |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | yesterday | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | yesterday | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 274k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |