Harness Evolve

What it does

When your agent uses it

• Tasks that involve Architecture decision records
• Tasks that involve Bioinformatics

Example prompts

• “/harness-evolve”

Requirements

• Pre-approved tools (allowed-tools): Bash

Workflow steps

6 steps, taken from the first numbered list in SKILL.md.

1. 1Validate args (--repo exists, caps on --generations ≤ 50, --children
2. 2Without --confirm: print plan + exit 0 (mirrors harness-mint safety
3. 3With --confirm: run metaharness-darwin evolve ... from the installed
4. 4Compute timeout from generations × children × per-variant (per-variant
5. 5Honor upstream exit code 99 — propagate as "safety-disqualified", do not
6. 6Optional --alert-on-no-improvement: exit 1 when champion ≤ parent.

What it can do on your machine

Read from SKILL.md and the folder at commit de590e1. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves these tools, so the agent can use them without asking each time:

  • Bash

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  No scripts in the folder and no shell commands in SKILL.md (its code samples are json).

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Harness Evolve loads about 1.6k tokens when it runs. Until then it costs about 114 tokens; SKILL.md has 601 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~114

When it runs · the whole SKILL.md, loaded when a task matches

~1.6k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

• NotePre-approves every shell command (allowed-tools: Bash)SKILL.md

  allowed-tools: Bash

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from ruvnet/ruflo at commit de590e1, republished under its MIT licence (© ruvnet). 601 words, ~1,578 tokens.

name

harness-evolve

description

Run `@metaharness/darwin evolve <repo>` to mutate a harness's seven policy surfaces (planner/contextBuilder/reviewer/retryPolicy/toolPolicy/memoryPolicy/scorePolicy), sandbox-score each variant, and promote only measured wins. The model is frozen; the harness evolves. Closes the loop ADR-150 opens (score+genome describe; evolve changes). Degrades gracefully when @metaharness/darwin is absent (ADR-150 + ADR-153 architectural constraints).

allowed-tools

Bash

argument-hint

--repo <path> [--generations 3] [--children 3] [--concurrency 2] [--sandbox real|mock|agent] [--selection pareto|quality-diversity|...] [--mutator…

.metaharness/
  archive.json         # full lineage tree (sampling next gen draws from this)
  lineage.json         # parent→child edges only
  variants/<id>/       # per-variant code (kept for audit)
  runs/<id>/           # per-variant sandbox test output
  reports/winner.json  # final champion + score delta vs parent

"diagnosis": {
  "available": true,
  "scope": "losing-variants",
  "variants": [
    { "id": "g1_v0", "transcripts": 2,
      "failureClasses": { "exploration-loop": 1, "edit-mechanics": 1 },
      "dominantClass": "exploration-loop" }
  ],
  "totals": { "exploration-loop": 1, "edit-mechanics": 1 }
}

© ruvnet, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in plugins/ruflo-metaharness/skills/harness-evolve of ruvnet/ruflo.

Open the folder on GitHubat commit de590e1