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 skills add jmagly/aiwg --skill execute-feedbackgit clone --depth 1 https://github.com/jmagly/aiwgWrote 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/skills/jmagly/aiwg/execute-feedback)<a href="https://agentmods.dev/skills/jmagly/aiwg/execute-feedback"><img src="https://agentmods.dev/badge/skills/jmagly/aiwg/execute-feedback/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/jmagly/aiwg/execute-feedback"><img src="https://agentmods.dev/badge/skills/jmagly/aiwg/execute-feedback.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00012 | $0.00645 |
| Opus 5.5 | $0.00005 | $0.00258 |
| Sonnet 5.5 | $0.00002 | $0.00129 |
| Haiku 4.5 | $0.00001 | $0.00064 |
Grade A, and why
execute-feedback 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 1mo 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 — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Execute Feedback Command
Run executable feedback loop on generated code: execute tests, analyze failures, fix, and retry.
Instructions
When invoked, perform the executable feedback loop per REF-013 MetaGPT:
-
Identify Target
- Load the specified file or recently modified code files
- Determine test framework (jest, pytest, cargo test, go test, etc.)
- Find existing tests or generate test stubs if none exist
-
Execute Tests
- Run the specified test command (or auto-detect)
- Capture full output (stdout, stderr, exit code)
- Parse test results: passed, failed, errors, skipped
-
Analyze Failures
- For each failing test:
- Extract error type and message
- Identify root cause (null check, type error, logic error, etc.)
- Map to source code location
- Check debug memory for similar past failures
- For each failing test:
-
Apply Fixes
- Generate targeted fix based on root cause analysis
- Apply fix to source code
- Increment attempt counter
-
Re-Execute
- Run tests again after fix
- Compare results to previous attempt
- If all pass: record success in debug memory, return
- If still failing: repeat from step 3
-
Escalate if Needed
- After max attempts (default: 3), escalate to human
- Include: all test results, failure analyses, fix attempts
- Save debug memory session
-
Update Debug Memory
- Record execution session in
.aiwg/ralph/debug-memory/sessions/ - Extract learned patterns to
.aiwg/ralph/debug-memory/patterns/ - Update success metrics
- Record execution session in
Arguments
[file-path]- Source file to test (default: recently modified files)--test-command [cmd]- Test command to run (default: auto-detect)--max-attempts [n]- Maximum fix attempts (default: 3)--coverage [%]- Minimum coverage target (default: 80)--no-fix- Run tests only, report without fixing--verbose- Show full test output
References
- @$AIWG_ROOT/agentic/code/frameworks/sdlc-complete/rules/executable-feedback.md - Executable feedback rules
- @$AIWG_ROOT/agentic/code/addons/agent-loop/docs/executable-feedback-guide.md - Implementation guide
- @$AIWG_ROOT/agentic/code/addons/agent-loop/schemas/debug-memory.yaml - Debug memory schema
- @$AIWG_ROOT/agentic/code/frameworks/sdlc-complete/schemas/flows/executable-feedback.yaml - Workflow schema
- @.aiwg/research/findings/REF-013-metagpt.md - Research foundation
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.
- 1mo ago Changed 8ca7b10747bc
- 1mo ago First seen · 72 lines · 12 tokens per session scan A 45a332467383
execute-feedback is a skill published in the GitHub repository jmagly/aiwg (224 stars, last pushed today), licensed MIT. It adds 12 tokens to every session and 645 once invoked, about $0.0000 per session on Opus 5.5. 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 skills, from other repositories
attest
Verifying spec compliance: extracts ACs from specs, adversarially checks conformance, generates BDD scenarios and traceability matrices. Use when impl must be proven to match a PRD/SRS/AC.
matrix
Controlling combinatorial explosion across multi-dimensional axes: minimum coverage sets, execution plans, test/deploy/UX/risk prioritization. Use when scoping multi-axis combinations.
siege
Verifying system resilience via load testing, contract testing, chaos engineering, and mutation testing. Use for limit verification, non-functional testing, or reliability validation.
probe
Integrating OWASP ZAP/Burp Suite/Nuclei, planning penetration tests, executing DAST, and scanning for vulnerabilities. For runtime vulnerability validation. Complements Sentinel static analysis.
voyager
Authoring web and native E2E tests, including Playwright, Appium, XCUITest, device farms, visual regression, and App Store screenshot pipelines. Not for unit/load tests.
radar
Adding edge-case tests, repairing flaky tests, and improving coverage. Use when test gaps need filling or regressions need guarding. Supports JS/TS, Python, Go, Rust, and Java.