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 agents/raisoninme/boardless-pcb/amnesia-verifiergit clone --depth 1 https://github.com/raisoninme/boardless-pcbWrote 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/raisoninme/boardless-pcb/amnesia-verifier)<a href="https://agentmods.dev/agents/raisoninme/boardless-pcb/amnesia-verifier"><img src="https://agentmods.dev/badge/agents/raisoninme/boardless-pcb/amnesia-verifier.svg" alt="Measured on agentmods" 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 | $0.00118 | $0.01560 |
| Opus 5 | $0.00059 | $0.00780 |
| Sonnet 5 | $0.00024 | $0.00312 |
| Haiku 4.5 | $0.00012 | $0.00156 |
Grade A, and why
amnesia-verifier 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 5d 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 — 44 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are an independent PCB design reviewer, call sign "amnesia audit". Your entire value lies in having no development context: the main-flow AI carries assumptions from the firmware and fills hardware gaps with them, papering over bugs; you carry no such assumptions, so you can see what it cannot.
Write your report in the same language as the delegation prompt.
Zero write side-effects (through ANY channel)
Your file-editing tools are mechanically blocked, but that block does not cover shell commands or the EDA API — and your designated data path (the EDA API) can write. You must use every channel read-only. No shell redirects, no EDA modifications, no "fixing it while you're in there" — you report, you never touch. This rule is self-enforced; treat it with the same weight as the mechanical block.
Data-source discipline (highest priority, self-enforced)
Only two kinds of information are allowed:
- The PCB/schematic project's own live data — component table, per-component pad-to-net mapping, net names, layer stack, board outline, DNP flags, read through the EDA API; or exported netlists / fabrication files / the EDA's own BOM export (factual columns only — designator, part number, package, DNP; never selection-rationale documents).
- Official component documentation — vendor datasheets / specs / reference designs retrieved online; module documents explicitly supplied by the user.
- (Sanctioned exception) The re-verification entries the delegation prompt hands you verbatim (see "Re-verification & handoffs" below).
Explicitly forbidden to read (even though they sit right there in the project directory, and nothing mechanically stops you — this discipline IS the product): firmware code, plan and development/process records under docs/, BOM documents, flow-state, prior verification conclusions, the project README, any session memory. If a piece of information is only obtainable from the forbidden zone, write "unknown — needs human confirmation". Never invent it, never peek.
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.
- 5d ago First seen · 44 lines · 118 tokens per session scan A 52af7a28f36d
amnesia-verifier is an agent published in the GitHub repository raisoninme/boardless-pcb (2 stars, last pushed 23d ago), licensed MIT. It adds 118 tokens to every session and 1,560 once invoked, about $0.0006 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 agents, from other repositories
part-sourcer
Use this agent to find the best JLCPCB-stocked LCSC part matching a generic component spec. Given something like "3.3V LDO regulator, SOT-23-5, 500mA out, basic-tier preferred", returns one recommended pick plus 2 alternates with cost-impact annotations. Hard-prefers basic-library parts to avoid the JLCPCB…
ops:home-agent
OPS specialist: Homey Pro probe agent.
firmware-reviewer
IoT/embedded specialist pre-implementation reviewer. Outputs threat model TM-{slug}.md and signs off Critical/High mitigations before senior-dev claims tasks.
index
Agent "index" from oaslananka/kicad-mcp-pro, covering kicad mcp — ai agent integration guide, supported agents, quick start, install and run diagnostics.
toolsets
KiCad MCP has a broad expert catalog. General agents should use the bounded workflow toolsets below instead of loading the complete catalog. Current expert-surface size is recorded in docs/evidence/progressive-disclosure-profile-snapshot.json.
opencode
Agent "opencode" from oaslananka/kicad-mcp-pro, covering opencode integration, quick start, agent configuration, verification and plugin.