harness
Skill2 starsChanged 8 months agoMITView on GitHub
Query daemon for harness engineering improvements, failures, and timeline — then apply them to improve the codebase for agents
- Sends data out
37 other projects use a version of this.
Adopted by
Most recently taken up by
- AshrafDurjoy/genai-shared-patternstheir own version, first seen Nov 16, 2026
- DavidTeam2008/harness-engineering-kittheir own version, first seen Nov 15, 2026
- manhquydev/flow-skilltheir own version, first seen Nov 13, 2026
- hanzhangzzz/agent-skills-zhtheir own version, first seen Oct 29, 2026
- anlmalabs/skillstheir own version, first seen Oct 25, 2026
- gutomec/nirvana-os-enginetheir own version, first seen Oct 13, 2026
Used with
Projects with it also use retro, verify and 3 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
AWSSkillretro
Conduct a retrospective on a coding session.
Also in 3 projects that use this
Microsoft AzureSkillverify
End-to-end app testing pipeline for browser-based web apps on cloud browsers via Azure Playwright Workspaces: verify the live app, author tests for working flows, heal failing tests, then re-run with video and produce one consolidated report. Three phases (app verification, test authoring, test healing) plus a scoped final sanity run; each phase emits a sub-report. Use when asked to "verify my app", "make sure my app works and has good test coverage", "do a full QA pass", "add tests where I'm missing coverage", "fix failing tests and add new ones", "smoke test before deploy", "QA my deployment", or any combined verify-and-test ask. Requires Azure CLI (`az login`) and a Playwright Workspace resource ID — bootstraps configs on first invocation. Browser-based web apps with Playwright only — not for unit/API tests, native mobile/desktop, load/perf, or full accessibility audits.
Also in 3 projects that use this
GoogleSkillcode-review
Adversarial systems, concurrency, memory-model, and architecture code review protocol. Use this skill whenever asked to review code, PRs, diffs, or branches in Filament.
Also in 4 projects that use this
GoogleSkillgrill-me
A relentless interview to sharpen a plan or design.
Also in 3 projects that use this
SentrySkillskill-creator
Create new agent skills following the Agent Skills specification. Use when asked to "create a skill", "add a new skill", "write a skill", "make a skill", "build a skill", or scaffold a new skill with SKILL.md. Guides through requirements, writing, registration, and verification.
Also in 3 projects that use this
Other versions37 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 37
- revfactory/harness★ 9.1ksame name
- cexll/myclaude★ 2.8ksame name
- Simpleyyt/ai-manus★ 1.6ksame name
- BankrBot/openclaw-skills★ 1.2ksame name
- BankrBot/skills★ 1.2ksame name
- bam-bam-2/solo-skills★ 366same name
- serejaris/personal-corp-os★ 244same name
- serejaris/ris-claude-code★ 243same name
- SensorsIot/Embedded-AI-Harness★ 235same name
- skillsynchq/txcript★ 208same name
27 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
12 rules. 0 are common, 8 are distinctive, and 4 are very specific to this project.
0
Common
Most of the 24 edited versions we read keep these.
8
Distinctive
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 19
If no active sessions, get the most recent:
- Distinctive: line 52
Look for patterns: repeated failure types, specific tools that fail often, permission issues.
- Very specific: line 68
Hook suggestions: Add to .claude/settings.json under the appropriate hook event
- Very specific: line 69
Skill suggestions: Create a new .claude/skills/<name>/SKILL.md file
- Very specific: line 70
Context suggestions: Update CLAUDE.md or create docs in docs/
- Distinctive: line 71
Architecture suggestions: Add lints, structural tests, or hook enforcement
- Distinctive: line 72
Subagent suggestions: Update CLAUDE.md with team strategy guidance
- Distinctive: line 73
Tool suggestions: Add MCP server configuration or create utility scripts
- Distinctive: line 75
Always ask the user before applying changes.
- Distinctive: line 79
After applying improvements, verify they work:
- Very specific: line 106
If the API returns errors, check if the server is running: curl -s localhost:3000/api/agent | jq
- Distinctive: line 108
Always preserve existing hooks when adding new ones — merge, don't replace
“Common” is measured: kept by at least 6 in 10 of the 24 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Harness Engineering
You are performing a harness engineering review. The Claude daemon is running at localhost:3000 and has been monitoring Claude Code sessions — capturing events, analyzing failures, and generating improvement suggestions.
Your job is to query the daemon API, understand what's working and what's not, and make concrete improvements to the codebase's agent harness.
Step 1: Discover Sessions
Find the most recent or active session:
curl -s localhost:3000/api/agent/sessions?status=active | jq '.data[0]'If no active sessions, get the most recent:
curl -s localhost:3000/api/agent/sessions?limit=1 | jq '.data[0]'Save the id field — you'll need it for all subsequent queries.
Step 2: Get Harness Improvements
Query the improvement suggestions:
curl -s "localhost:3000/api/agent/improvements?sessionId=SESSION_ID" | jq '.data'These are categorized by area:
- hooks — Pre/post tool hooks for enforcement, auto-formatting, blocking dangerous patterns
- skills — Custom slash commands for repeatable workflows
- subagents — Agent teams for parallel work, specialized roles
- tools — MCP servers, custom tools, observability integrations
- context — CLAUDE.md, docs/, design docs, architecture maps
- architecture — Layer boundaries, dependency rules, naming conventions
- legibility — Structured logging, error messages with remediation, self-documenting APIs
Step 3: Get Failures
Query the detected failures:
curl -s "localhost:3000/api/agent/failures?sessionId=SESSION_ID" | jq '.data'Look for patterns: repeated failure types, specific tools that fail often, permission issues.
Step 4: Present Findings
Summarize what you found, grouped by category:
- Critical (severity: high) — Address these first
- Friction (severity: medium) — These slow agents down
- Nice-to-have (severity: low) — Polish items
For each suggestion that has a config field, show the ready-to-use configuration.
Step 5: Apply Improvements
For each suggestion, offer to implement it:
- Hook suggestions: Add to
.claude/settings.jsonunder the appropriate hook event - Skill suggestions: Create a new
.claude/skills/<name>/SKILL.mdfile - Context suggestions: Update
CLAUDE.mdor create docs indocs/ - Architecture suggestions: Add lints, structural tests, or hook enforcement
- Subagent suggestions: Update
CLAUDE.mdwith team strategy guidance - Tool suggestions: Add MCP server configuration or create utility scripts
Always ask the user before applying changes. Show what you'll change and get confirmation.
Step 6: Verify
After applying improvements, verify they work:
- For hooks: trigger the relevant tool use and check the hook fires
- For skills: run the skill with
/skillname - For context: read the file and verify it's correct
- For architecture: run
npx tsc --noEmitand linters
Advanced: GraphQL Queries
For complex queries, use the GraphQL endpoint:
curl -s -X POST localhost:3000/api/agent/graphql \
-H 'Content-Type: application/json' \
-d '{"query": "{ improvements(sessionId: \"SESSION_ID\", severity: HIGH) { area problem suggestion config effort } }"}'Multi-level timeline:
curl -s -X POST localhost:3000/api/agent/graphql \
-H 'Content-Type: application/json' \
-d '{"query": "{ timeline(sessionId: \"SESSION_ID\") { plans(level: 0) { name phase tasks { name status } } } }"}'Important
- The daemon must be running at
localhost:3000for this skill to work - If the API returns errors, check if the server is running:
curl -s localhost:3000/api/agent | jq - Prioritize high-severity improvements with trivial/small effort first — maximum impact, minimum cost
- Always preserve existing hooks when adding new ones — merge, don't replace
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

