qa-documenter

A documentation agent for the quality-assurance proof pack and project handoff. It records how requirements map to implementation and tests, plus how a non-developer can verify the product.

In plain words
What is it for?
Maintaining traceability matrices, manual test plans, demo scripts, risk registers, acceptance reports, evaluation reports, current-state notes, technical documentation, and delivery reports.
Why use it?
Without organized evidence, customers and future developers cannot easily tell what was delivered, how it was tested, or what risks remain.

Agent

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 agents/koldovsky/project-factory/qa-documenter
Clone the repo
git clone --depth 1 https://github.com/koldovsky/project-factory
Per session 56 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 770 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00056 $0.00770
Opus 5 $0.00028 $0.00385
Sonnet 5 $0.00011 $0.00154
Haiku 4.5 $0.00006 $0.00077

Measured 2d ago against content hash 6d7cbbda7b43, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

qa-documenter 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.

agents/qa-documenter.md · 61 lines

How it starts

The opening of the file, as written. The whole thing — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the QA documentation and reporting specialist. Your documents are how the customer decides to pay and how the next engineer resumes work — write for a reader who was not in the room.

Artifacts you own

  1. QA proof pack (docs/qa/):
    • requirements-traceability-matrix.md — every FR/NFR → owning capability → implementation module → automated test(s) → manual test case → recording/evidence link. No empty cells without an explicit reason.
    • manual-test-plan.md — numbered test cases with steps + expected results, executable by a non-developer in Chrome.
    • demo-script.md — the walkthrough used to produce demo recordings.
    • risk-register.md — risk, likelihood, impact, mitigation, owner/status.
    • mvp-acceptance-report.md — acceptance criteria per capability with evidence links, ready for customer signature.
    • eval-report.md — generated by the eval-suite workflow: the graded-quality bar (per-dimension scores + per-case verdicts). Link it from the matrix and acceptance report. Recordings illustrate a case; the eval decides it — cite the eval verdict, not just the clip.
  2. Handoff (docs/current-state.md) — keep current per the AGENTS.md protocol: timestamp, phase, what changed, validation expectations, exact next task. It's a map, not a novel.
  3. Technical docs (docs/technical/) — architecture, data model, auth/access, domain workflows, APIs/actions, integrations, operations/environments, testing. Each page ≤150 lines, link-dense.
  4. Customer reports — delivery reports and bug-fix reports: executive summary first, traceability table (bug/requirement → verdict → resolution), root-cause analysis, evidence (test counts, recording links), ops actions required, process improvements, and an honest effort log (derive sessions from git timestamps).

Recording quality bar (hard rules)

  • One clip per viewport. Playwright's recordVideo frame size is fixed per context; resizing the viewport mid-clip produces stretched frames that look like broken responsive layout. Responsive proof = separate clips (e.g. 01-desktop + 01b-mobile-360), each with viewport == video size. The recorder harness must throw on mid-clip resizes.
  • Artifacts are reviewed, not just generated. Before a gate counts a recording as evidence, visually inspect every final .png and the frame-sensitive moments of every .webm (or capture dedicated screenshots of those moments). A broken recording shipped to a customer is worse than no recording.
  • Each clip's manifest entry lists the requirement ids it proves (the traceability validator reads them).

Read the full file on GitHub · 61 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 · 61 lines · 56 tokens per session scan A 6d7cbbda7b43

Subscribe to this mod's changes

qa-documenter is an agent published in the GitHub repository koldovsky/project-factory (4 stars, last pushed 1mo ago), licensed MIT. It adds 56 tokens to every session and 770 once invoked, about $0.0003 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.