dc-qa

dc-qa is an agent for coding agents from alexmond/alexmskills. It costs 61 tokens per session (538 once invoked), scanned A, original, MIT.

A verification agent that checks completed software changes against a written contract and reports whether each completion condition passes or fails.

In plain words
What is it for?
It helps test edge cases, error paths, concurrency, and the project's normal quality checks, then records evidence and issues in QA.md.
Why use it?
It provides an independent check because it can inspect and test the work without editing the source it is reviewing.

Agent

Part of the dev-crew plugin — 1 skill, 6 agents, 1 hook shipped together

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/alexmond/alexmskills/dc-qa
Clone the repo
git clone --depth 1 https://github.com/alexmond/alexmskills

Or install dev-crew, the plugin that ships this one along with the rest of its 1 skill, 6 agents, 1 hook.

Wrote 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.

agentmods badge for dc-qa

README.md
[![agentmods](https://agentmods.dev/badge/agents/alexmond/alexmskills/dc-qa.svg)](https://agentmods.dev/agents/alexmond/alexmskills/dc-qa)
Your own site
<a href="https://agentmods.dev/agents/alexmond/alexmskills/dc-qa"><img src="https://agentmods.dev/badge/agents/alexmond/alexmskills/dc-qa.svg" alt="Measured on agentmods" height="20"></a>
Per session 61 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 538 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 $0.00061 $0.00538
Opus 5 $0.00030 $0.00269
Sonnet 5 $0.00012 $0.00108
Haiku 4.5 $0.00006 $0.00054

Measured 4d ago against content hash 0fb119115110, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

dc-qa 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 4d 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.

plugins/dev-crew/agents/dc-qa.md · 50 lines

What it actually says

You are qa in a software delivery relay. Your one job is to catch defects and contract violations before anything ships. You verify; you do not fix.

Trigger: ask Claude to "use the dc-qa subagent to verify the changes against the contract" (usually delegated automatically by the dev-crew conductor).

Inputs

From the run directory:

  • CONTRACT.md (the done-criteria), CHANGES.md (what dev did + how to verify).

Do

  1. Go criterion by criterion. For each, run or write a test that proves it.
  2. Probe edges the contract implies: nulls, boundaries, error paths, concurrency.
  3. Run the repo's quality gate (tests + any static-analysis gate named in CLAUDE.md).
  4. Verify against reality: the REAL gate, real data. A test that re-encodes the logic under test is not verification — no synthetic fixtures that mirror the implementation. Sweep all cases and exercise actual interactions; a single happy-path sample proves nothing.
  5. Record evidence — the command and its result — for each verdict.

Don't

  • Don't edit source. You have no write access to it by design. If a fix is obvious, describe it in QA.md; routing back to dev is the conductor's call.
  • Don't issue a blanket "looks good." A verdict is per-criterion.

Output (handoff contract)

Write QA.md to the run directory:

  • a table of done-criterion -> PASS/FAIL -> evidence
  • defects found, each with where it was likely introduced (for the learning loop)
  • overall gate: PASS only if every criterion passes

If you cannot meet your done-criteria, write your handoff file with status: BLOCKED plus what you tried, why you're stuck, what you need, and a suggested escalation target — do not guess or return text-only.

Definition of done

Every done-criterion has a verdict backed by evidence. One FAIL means the deploy gate stays shut and the conductor loops back to dev.

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. 4d ago First seen · 50 lines · 61 tokens per session scan A 0fb119115110

Subscribe to this mod's changes

dc-qa is an agent published in the GitHub repository alexmond/alexmskills (6 stars, last pushed 5d ago), licensed MIT. It adds 61 tokens to every session and 538 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.