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

harness-reflector

Skill1 starsChanged 20 days agoApache-2.0View on GitHub

Use when the user asks to "learn from this task", "save this workflow as a skill", "update AGENTS.md with lessons", or reflect on a completed debugging or implementation session. Produces verified, scoped rules and reusable skills instead of speculative advice.

  • Installs packages

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    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:

    Run this protocol after a substantial task when the user asks for deliberate reflection, or when the automatic Stop cycle did not capture …

    line 12
  2. Very specific:

    If the transcript is not available, create a skill with skills install from a local folder, or add a concise directive to AGENTS.md …

    line 62

≈ “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 Deep Reflector

Run this protocol after a substantial task when the user asks for deliberate reflection, or when the automatic Stop cycle did not capture something valuable.

1. Gather evidence

  • What was the user's actual goal?
  • Which commands, edits, or tool calls changed the outcome?
  • What failed? Include the exact error, not a paraphrase.
  • What fixed it? Only count a fix that was followed by a successful command, passing test, or otherwise observable verification.
  • Was the failure transient (timeout, rate limit, flaky network)? If so, do not turn it into a permanent rule.

2. Write bounded lessons

Good rule shape:

pytest failed with ModuleNotFoundError: No module named 'pytest'. Verified fix: run python -m pip install pytest in the active environment.

Bad rule shape:

Command failed. Try again or adjust parameters.

Reject rules containing timestamps, UUIDs, exit codes alone, truncated output, or no actionable recovery.

3. Decide scope

  • Workspace: project commands, paths, package manager, migrations, tests, local conventions. Write to the workspace AGENTS.md.
  • Global: cross-project environment quirks such as PowerShell quoting, Python UTF-8 behavior, or hook cwd. Never copy project paths or commands into the global file.
  • Skill: a repeatable multi-step procedure, not a single command and not a one-off incident. Require at least three meaningful steps, or two steps plus a verification step.

4. Register through Harness

# Re-analyse the current session and update workspace rules/skills
bin\harness.cmd reflect --workspace . --force

# Inspect what was learned
bin\harness.cmd status
bin\harness.cmd skills
bin\harness.cmd rules show --target workspace

If the transcript is not available, create a skill with skills install from a local folder, or add a concise directive to AGENTS.md manually. Never invent a verified recovery.

5. Verify

  • bin\harness.cmd skills validate
  • bin\harness.cmd rules sanitize --dry-run
  • Confirm the rule/skill contains no credentials or project-irrelevant noise.
  • Confirm the managed block stays under the 20 KB budget and the file stays under Antigravity's 24 KB cap.

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