Skip to content
Back to skills

Os Evolution Verifier

ASecurity

Verifies that os-architect actually causes evolution — not just words. Dispatches os-architect in single-shot simulation mode for a given test scenario, then checks for real artifact presence (new files, HANDOFF_BLOCK, plan files). Reports PASS / FAIL with grep evidence. Accumulates results into a test report. Use after any changes to os-architect, os-evolution-planner, or improvement-intake-agent.

  • 7 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 2, 2026
ai-agentspythonbash

Security analysis

A100/100

Pro scans all 5 files and shows the line behind each finding

Scanned October 3, 2026

npx -y skills add richfrem/agent-plugins-skills --skill os-evolution-verifier --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Os Evolution Verifier?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Os Evolution Verifier
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/richfrem-os-evolution-verifier/badge)](https://www.skillsdirectory.com/skills/richfrem-os-evolution-verifier)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: os-evolution-verifier
plugin: agent-agentic-os
description: >
  Verifies that os-architect actually causes evolution — not just words.
  Dispatches os-architect in single-shot simulation mode for a given test scenario,
  then checks for real artifact presence (new files, HANDOFF_BLOCK, plan files).
  Reports PASS / FAIL with grep evidence. Accumulates results into a test report.
  Use after any changes to os-architect, os-evolution-planner, or improvement-intake-agent.
argument-hint: "[test-scenario-file | all]"
tools: ["Bash", "Read", "Write"]
---

# Evolution Verifier (`os-evolution-verifier`)

Verifies that evolution agents cause real changes by checking for artifact presence (new files, handoff blocks, plan files) rather than relying on transcript text.

## Contents

- [Critical Constraints](#critical-constraints)
- [Quick start](#quick-start)
- [Workflow](#workflow)
- [Artifact Verification Matrix](#artifact-verification-matrix)
- [Verification](#verification)
- [References](#references)

## Critical Constraints

1. **Physical Artifact Verification**: Evolution is proven strictly by physical file existence and content presence, never conversational transcripts.
2. **Binary Pass/Fail Contract**: A run only passes if expected files exist and required handoff block fields are complete.
3. **Persist Results**: Always record verification outcomes to `os-experiment-log` before `temp/` is cleared.

## Quick start

Run verification on a single evolution scenario:

```bash
python3 scripts/run_evolution_scenario.py --scenario <path-to-scenario.json>
```

## Workflow

1. **Resolve Scenarios**: Locate target test scenario or scan `temp/os-evolution-verifier/scenarios/*.json`.
2. **Dispatch Simulation**: Dispatch `os-architect` in single-shot non-interactive simulation mode via subagent.
3. **Verify Artifacts**: Check physical existence of target files and validate `HANDOFF_BLOCK` integrity.
4. **Compile Report**: Append PASS/FAIL assessment block to test report.
5. **Persist Findings**: Invoke `os-experiment-log` to archive test outcomes permanently.

## Artifact Verification Matrix

| Evolution Type | Verification Target |
|---|---|
| Gap Fill (Path C) | `SKILL.md` present at expected spoke path |
| Update (Path B) | `tasks/todo/<slug>-plan.md` and prompt written |
| No-Op (Path A+) | No unexpected files; `HANDOFF_BLOCK` complete |
| Lab Setup | `run-config.json` written with valid configuration |

## Verification

Confirm test report is compiled and recorded:

```bash
test -f "temp/os-evolution-verifier/test-report.md" && echo "Report verified"
```

## References

- [detailed-reference.md](references/detailed-reference.md) — Exact dispatch commands, EVOLUTION_VERIFICATION schema, and failure trees.
- [acceptance-criteria.md](references/acceptance-criteria.md) — Pass/fail grading criteria and scenario formats.
- [fallback-tree.md](references/fallback-tree.md) — Remediation pathways when simulation or verification fails.

Files in this skill

  • SKILL.md13.4 KB
  • evals/evals.json1.1 KB
  • references/acceptance-criteria.md42 B
  • references/cheapest_models.md38 B
  • scripts/experiment_log.py34 B

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.