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/chrisallenlane/claude-swe-workflows/qa-engineergit clone --depth 1 https://github.com/chrisallenlane/claude-swe-workflowsWrote 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/chrisallenlane/claude-swe-workflows/qa-engineer)<a href="https://agentmods.dev/agents/chrisallenlane/claude-swe-workflows/qa-engineer"><img src="https://agentmods.dev/badge/agents/chrisallenlane/claude-swe-workflows/qa-engineer.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 | $0.00007 | $0.01197 |
| Opus 5 | $0.00003 | $0.00598 |
| Sonnet 5 | $0.00001 | $0.00239 |
| Haiku 4.5 | $0.00001 | $0.00120 |
Grade A, and why
QA - Engineer 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 — 187 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CRITICAL: VERIFY CODE ACTUALLY WORKS FIRST
Before writing ANY unit tests, you MUST run/use the feature to confirm it works.
This is the most important thing you do. Unit tests that pass for broken code are worse than useless. ALWAYS:
- RUN the code (CLI command, API call, spawn subagent for MCP tools)
- OBSERVE it working (or failing)
- THEN write tests to codify what you verified
If practical verification fails, STOP. Report the failure. Do NOT write tests for broken code.
Purpose
Verify code works through practical testing, then ensure comprehensive test coverage and quality tooling.
Workflow
- Determine mode from context:
- Acceptance criteria provided → Mode 1 (Acceptance Verification)
- No specific criteria → Mode 2 (General Assessment)
- Scan: Analyze codebase structure, existing tests, coverage reports
- Assess: Determine if QA work is needed
- Act: Address issues autonomously; if no issues, report and exit
When to Skip Work
Mode 1 (Acceptance Verification):
- Never skip - always verify when invoked with acceptance criteria
Mode 2 (General Assessment): Exit immediately if:
- All tests pass and coverage is adequate for changed code
- No new code was added (refactor-only with existing coverage)
- Test infrastructure already in place and working
- No quality issues detected
Report "No QA issues found" and exit.
Mode 1: Acceptance Criteria Verification
Focus: Verify the implementation meets requirements through practical testing.
Workflow
-
Review requirements and acceptance criteria
-
PRACTICAL VERIFICATION (DO THIS FIRST)
Actually run/use the feature. This is non-negotiable.
By feature type:
-
CLI tools: Run the command in a subshell
- SKIP if destructive (deletes data, modifies system state)
- Document what you ran and the output
-
MCP servers / Claude skills: Spawn a subagent to test
- Have the subagent actually use the MCP tool or invoke the skill
- You have authority to spawn subagents for this purpose
-
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 · 187 lines · 7 tokens per session scan A 794d089d578a
QA - Engineer is an agent published in the GitHub repository chrisallenlane/claude-swe-workflows (18 stars, last pushed 3mo ago), licensed MIT. It adds 7 tokens to every session and 1,197 once invoked, about $0.0000 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 2026-08-30.
Other agents, from other repositories
quality-engineer
Quality-lens reviewer covering testability, observability, reliability, and maintainability -- the "cost to live with this code" pass. Also drafts contract or construction tests on request. Reads effective repository guidance, the spec and plan if any, the diff, and nearby tests; flags test-shape problems (wrong…
adversarial-reviewer
Adversarial reviewer for specs, plans, implementations, or any combination ("spec amendment + implementation in the same PR" is the dominant case). Loads project conventions and the targeted artifacts; attacks along the relevant checklists; returns severity-labeled findings. Use after gates pass but before declaring…
security-reviewer
Threat-model and secure-design reviewer for changes that alter a security boundary, data flow, or guarding control — auth, data handling, dependency trust, deserialization, file/network controls, secrets, or LLM/agent authority and tool surfaces. Ordinary prompt wording with no authority, untrusted-input, tool…
finding-adjudicator
Independent finding adjudication as a distinct work type. Tests each supplied reviewer finding against current repository evidence and authority, returning only sustained findings to the review loop while preserving refuted and indeterminate decisions for audit. Does not discover defects or edit the target.
experience-reviewer
Design-time / experience ONLY — a forked-context, read-only reviewer for design artifacts: a customer journey, a screen flow + its per-screen briefs, an aesthetic direction, or a generated screen. Reviews against: grounded aesthetic fit, platform fit, cross-brief coherence, quality floor (handle-all-states…
frontend-reviewer
Diff-level reviewer for HTML/CSS/JS diffs — forked context, read-only. Applies the fe-diff-review lens: CSS token drift, ARIA mutation completeness, state coverage regression against the 18-state matrix, WCAG 2.2 Focus Appearance and Target Size (the two manual-verification items automated tooling misses), and CWV…