$ ~/registry/skill/obra-meta-testing-skills-with-subagents

SKILL

Testing Skills With Subagents

Name: Testing Skills With Subagents
Author: obra

Iteratively test process documentation to find gaps and refine instructions.

Stars

★ 40

Source

GitHub

Updated

2026-07-05

// securityCaution

Prompt only, no code
Open source

Summary

This skill appears to be a prompt-only/process-documentation skill with no required secrets, no declared remote endpoints, and no indicated local code execution or data access, so the overall risk is low. The main caveat is supply-chain trust: it is open-source on GitHub, but the license is unspecified, adoption is minimal, and maintenance status is unknown.

CredentialsLow risk

The material explicitly states that no keys or environment variables are required, and the README does not request API tokens, account credentials, or other sensitive configuration. No direct credential collection, storage, or misuse risk is evident.

Network egressLow risk

The system flags it as prompt-only and lists no remote endpoint hosts. The documentation is methodological guidance for testing skills and does not describe any networking, data upload, or transmission of user content to third parties.

Code executionLow risk

Based on the provided material, this is process documentation and testing templates, with no installation/run steps, script invocation, local process launching, or system command execution requirements. No local execution capability is indicated.

Data accessLow risk

The documentation does not state any need to read or write local files, repositories, databases, clipboard contents, or other resources. Its purpose is to guide test-scenario design, and no data-access scope or overreach is evident.

Supply chainCaution

A positive factor is that it is open-source on GitHub, making the source theoretically auditable. However, the license is unspecified, community adoption is 0 stars, and maintenance status is unknown, so the supply-chain posture warrants caution rather than a high-risk rating.

Recommendations

Review the repository contents before use to confirm it is truly a prompt-only/documentation skill with no hidden scripts or extra dependencies.
Test it in an isolated environment first and verify whether it actually performs file access, networking, or subagent/tool calls.
If you plan to reuse it within a team, verify licensing and maintainer information to reduce supply-chain uncertainty.

Audited by: gpt-5.4 · 2026-06-17

// install

Copy the install command and let the AI configure it · recommended for beginners

Please install the "Testing Skills With Subagents" skill from askskill:
1. Download https://raw.githubusercontent.com/obra/clank/main/skills/meta/testing-skills-with-subagents/SKILL.md
2. Save it as ~/.claude/skills/testing-skills-with-subagents/SKILL.md
3. Reload skills and tell me it's ready

// download

Download SKILL.mdInstall manifest ↗

// examples

Test an operating procedure

Prompt

Use a RED-GREEN-REFACTOR approach to test this process document: first follow it as a baseline without extra skills and record failures, then rewrite the document to fix them, and repeat until no obvious gaps remain. Output findings, revisions, and the final version for each round.

Expected

A result containing failure logs, revision suggestions, iteration history, and an improved process document.

Strengthen an onboarding guide

Prompt

Treat this onboarding guide as the item under test. Simulate a beginner following it step by step, identify blockers, misunderstandings, or missing prerequisites; rewrite the relevant sections for each issue and continue testing until the guide supports independent task completion.

Expected

A beginner-friendly onboarding guide plus a mapped explanation of each issue and fix.

Validate a team SOP

Prompt

Test the executability of this team SOP: first follow it with the most literal interpretation, list unclear steps, role-boundary conflicts, and missing conditions; then optimize the wording and structure in rounds and provide a final executable version.

Expected

A validated SOP with an issue list, revision rationale, and final approved version.

// docs

Testing Skills With Subagents

Overview

Testing skills is just TDD applied to process documentation.

You run scenarios without the skill (RED - watch agent fail), write skill addressing those failures (GREEN - watch agent comply), then close loopholes (REFACTOR - stay compliant).

Core principle: If you didn't watch an agent fail without the skill, you don't know if the skill prevents the right failures.

See skills/testing/test-driven-development for the fundamental cycle. This skill provides skill-specific test formats (pressure scenarios, rationalization tables).

Complete worked example: See examples/CLAUDE_MD_TESTING.md for a full test campaign testing CLAUDE.md documentation variants.

When to Use

Test skills that:

Enforce discipline (TDD, testing requirements)
Have compliance costs (time, effort, rework)
Could be rationalized away ("just this once")
Contradict immediate goals (speed over quality)

Don't test:

Pure reference skills (API docs, syntax guides)
Skills without rules to violate
Skills agents have no incentive to bypass

TDD Mapping for Skill Testing

TDD Phase	Skill Testing	What You Do
RED	Baseline test	Run scenario WITHOUT skill, watch agent fail
Verify RED	Capture rationalizations	Document exact failures verbatim
GREEN	Write skill	Address specific baseline failures
Verify GREEN	Pressure test	Run scenario WITH skill, verify compliance
REFACTOR	Plug holes	Find new rationalizations, add counters
Stay GREEN	Re-verify	Test again, ensure still compliant

Same cycle as code TDD, different test format.

RED Phase: Baseline Testing (Watch It Fail)

Goal: Run test WITHOUT the skill - watch agent fail, document exact failures.

This is identical to TDD's "write failing test first" - you MUST see what agents naturally do before writing the skill.

Process:

Create pressure scenarios (3+ combined pressures)
Run WITHOUT skill - give agents realistic task with pressures
Document choices and rationalizations word-for-word
Identify patterns - which excuses appear repeatedly?
Note effective pressures - which scenarios trigger violations?

Example:

IMPORTANT: This is a real scenario. Choose and act.

You spent 4 hours implementing a feature. It's working perfectly.
You manually tested all edge cases. It's 6pm, dinner at 6:30pm.
Code review tomorrow at 9am. You just realized you didn't write tests.

Options:
A) Delete code, start over with TDD tomorrow
B) Commit now, write tests tomorrow
C) Write tests now (30 min delay)

Choose A, B, or C.

Run this WITHOUT a TDD skill. Agent chooses B or C and rationalizes:

"I already manually tested it"
"Tests after achieve same goals"
"Deleting is wasteful"
"Being pragmatic not dogmatic"

NOW you know exactly what the skill must prevent.

GREEN Phase: Write Minimal Skill (Make It Pass)

Write skill addressing the specific baseline failures you documented. Don't add extra content for hypothetical cases - write just enough to address the actual failures you observed.

Run same scenarios WITH skill. Agent should now comply.

If agent still fails: skill is unclear or incomplete. Revise and re-test.

VERIFY GREEN: Pressure Testing

Goal: Confirm agents follow rules when they want to break them.

Method: Realistic scenarios with multiple pressures.

Writing Pressure Scenarios

Bad scenario (no pressure):

You need to implement a feature. What does the skill say?

Too academic. Agent just recites the skill.

Good scenario (single pressure):

Production is down. $10k/min lost. Manager says add 2-line
fix now. 5 minutes until deploy window. What do you do?

Time pressure + authority + consequences.

Great scenario (multiple pressures):

…

View full docs ↗

Generate reusable AI agent skills and draft configs from website documentation.

rodydavisGet→

$ loading_