agentleFS
Sign inStart free
Ai-assistant-KIRAAi-assistant-KIRA/antigravity-harness/skills/harness-control/SKILL.md

harness-control

Skill1 starsChanged 19 days agoApache-2.0View on GitHub

Use when the user asks about Harness², learned rules, memory, self-improvement, synthesized skills, or wants to inspect, reflect on, export, sanitize, or manage the Harness plugin. Covers harness status, reflect, skills, memory, rules, and doctor commands.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Prefer the wrapper installed with the plugin:

    line 18
  2. Distinctive:

    Never hand-edit the managed block in AGENTS.md.

    line 56
  3. Distinctive:

    Do not promote an unresolved error into a rule.

    line 58
  4. Distinctive:

    Keep workspace rules in the workspace AGENTS.md.

    line 60
  5. Distinctive:

    Never store credentials, tokens, private keys, or connection strings.

    line 62
  6. Very specific:

    If doctor reports an invalid skill, fix or remove it before the next session.

    line 64

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Harness² Control

Harness² learns from completed sessions: it records verified failure/recovery lessons, evolves workspace AGENTS.md rules, and synthesizes or refines reusable skills. Use this skill when the user wants to inspect or operate that system.

Commands

Prefer the wrapper installed with the plugin:

# Windows (from the plugin directory, or add bin\ to PATH)
bin\harness.cmd status
bin\harness.cmd skills
bin\harness.cmd memory search "pytest import"
bin\harness.cmd rules show --target both
bin\harness.cmd reflect --workspace . --force
# macOS / Linux
bin/harness status
bin/harness skills
bin/harness memory search "pytest import"
bin/harness rules show --target both
bin/harness reflect --workspace . --force

From a checkout, python scripts/harness_cli.py <command> works from any working directory.

Useful operations

  • status shows sessions, verified lessons, candidates, skills, and rule sizes.
  • skills list shows workspace, plugin, and user skills. skills show NAME prints one. skills validate checks frontmatter and credentials.
  • skills install SOURCE installs a local skill folder, owner/repo, GitHub tree URL, or raw SKILL.md. Use --scope workspace for project skills.
  • memory search QUERY searches lessons with FTS5 when available.
  • memory export PATH / memory import PATH moves the store between machines.
  • memory prune removes stale candidates and bounds the database.
  • rules show / rules sanitize inspects or repairs managed rule blocks.
  • doctor checks deployment, hooks, rule sizes, and invalid skills.

Rules of engagement

  1. Never hand-edit the managed block in AGENTS.md; Harness owns it. User text outside the block is preserved.
  2. Do not promote an unresolved error into a rule. Only a verified recovery is evidence.
  3. Keep workspace rules in the workspace AGENTS.md. Global rules are for environment and tooling guidance only.
  4. Never store credentials, tokens, private keys, or connection strings. The analyzer redacts before persistence, but do not rely on it as a safety net.
  5. If doctor reports an invalid skill, fix or remove it before the next session; invalid frontmatter is ignored by progressive disclosure.

What happens automatically

  • SessionStart injects a short briefing with relevant skills and lessons.
  • PostToolUse journals failures for diagnostics without blocking the agent.
  • Stop analyses the transcript, updates rules and skills, and records memory.
  • Re-running the same transcript is a no-op unless --force is passed.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.

Check your setup, free