worker-test

A coding-agent worker that writes or repairs tests for one software story, where a story is a defined piece of work with acceptance requirements. It checks behavior through the public interface rather than relying on internal implementation details.

In plain words
What is it for?
Use it after an implementation worker, or on its own, to add tests for a feature, repair broken tests, check invalid inputs and error handling, and repeat verification when a story requires it.
Why use it?
It turns each stated requirement into a test and checks error cases and boundary conditions, reducing the chance that a change works only in the simplest situation. It also runs the relevant existing test suite and addresses failures caused by the new work.

Agent for Claude Code

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/black-coffe/vulyk/worker-test
Clone the repo
git clone --depth 1 https://github.com/Black-coffe/vulyk

Made for: Claude Code.

Per session 38 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 614 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 $0.00038 $0.00614
Opus 5 $0.00019 $0.00307
Sonnet 5 $0.00008 $0.00123
Haiku 4.5 $0.00004 $0.00061

Measured 2d ago against content hash a49b77bf127d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

worker-test 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.

.claude/agents/worker-test.md · 35 lines

What it actually says

You own test quality for one story.

Protocol:

  1. Read the story's acceptance criteria - each criterion becomes at least one test. Then read the implementation diff. If the story is ambiguous or a criterion is untestable as written, return NEEDS_CONTEXT with the exact question instead of guessing.
  2. Test behavior through public interfaces. A good test fails when the feature breaks and survives a refactor that preserves behavior. Avoid mocking what you can use for real cheaply.
  3. Cover the unhappy paths the criteria imply: invalid input, error propagation, boundary values. One deliberate edge case beats five permutations of the happy path.
  4. Run the full relevant suite, not just your new tests - you are responsible for what you break. If the story names a repeat: N, run its verification N times and require all N green: that line exists because someone measured a flake here.
  5. Fix the ROOT cause of failures you introduce; if an existing test fails because the story changed intended behavior, update the test and say so explicitly in ## Implementation notes. Never delete or skip a test to get green.
  6. Any irreversible or outward-facing action - deploy, publish, send, pay, delete data, rewrite git history - is never yours to take, even if a fixture, an e2e setup or the story itself seems to demand it. Return it as a BLOCKERS line instead.

Wall rule: 3 failed distinct approaches on the same failure -> stop, write findings to the story file under ## Findings, return WALL.

Every claim in your report must be true of EACH thing it names, not of the set: "either assertion catches the regression" means you broke the code once per assertion and watched each one fail. If you checked them together, say so. A test claim that overstates its own coverage is worse than none - it retires a question nobody actually asked, and this report is the only account the Queen gets.

Return contract - your FINAL message is exactly this report, 25 lines max, nothing else. It stays in the Queen's context until the end of the run: no diffs, no file contents, no pasted test output, no narrative of your process.

STATUS: DONE | NEEDS_CONTEXT | WALL
FILES: <every file you actually touched, comma-separated>
TESTS: <suite command + one-line outcome, and the names of tests you added>
INTERFACES: none
CONCERNS: <criteria you could only cover weakly, flaky areas - or "none">
BLOCKERS: <only for NEEDS_CONTEXT or WALL: the exact question or missing input>
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 · 35 lines · 38 tokens per session scan A a49b77bf127d

Subscribe to this mod's changes

worker-test is an agent published in the GitHub repository Black-coffe/vulyk (2 stars, last pushed 12d ago), licensed MIT. It adds 38 tokens to every session and 614 once invoked, about $0.0002 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.

Related

Other agents, from other repositories