All skills
oaustegard avatar

/optimizing-skills

@59f45d6

Disciplined, validation-gated revision of an EXISTING skill so each edit is a measured improvement rather than a guess. Use when editing, revising, or tuning a skill that already exists and there is evidence it underperforms (observed failures, drift, complaints) — invoke by name, or have versioning-skills / creating-skill defer to it before applying edits. Not for authoring a brand-new skill from scratch (use creating-skill) or one-off prose.

Use this Skill: https://skilld.dev/gh/oaustegard/claude-skills/optimizing-skills

This session only. Nothing lands on disk.

README.md

≈209 tokens on demand. Your agent reads this file only when SKILL.md points to it.

optimizing-skills

Disciplined, validation-gated revision of an existing skill so each edit is a measured improvement rather than a guess. Use when editing, revising, or tuning a skill that already exists and there is evidence it underperforms (observed failures, drift, complaints) — invoke by name, or have versioning-skills / creating-skill defer to it before applying edits. Not for authoring a brand-new skill from scratch (use creating-skill) or one-off prose.

Distilled from SkillOpt (microsoft/SkillOpt): a held-out check set, a two-tier best/candidate validation gate (ship only on a strict improvement), bounded edits, failure-first reflection, impact ranking, a protected core, and remember()-backed cross-revision memory — the discipline kept, the training harness dropped.

Source: SKILL.md on GitHub

No alerts3mo3 checks · Risk SAFE
  • Gen Agent Trust Hub3mo

    The skill provides a disciplined framework for optimizing AI agent skills through evidence-based revision and validation. It leverages platform-provided tools for subagent execution and persistent memory. A minor security risk exists in the form of indirect prompt injection, as the skill processes untrusted test tasks and transcripts without explicit boundary markers or sanitization when passing them to subagents for evaluation and reflection.

  • Socket3mo

    No alerts

  • Snyk3mo

    Risk: LOW · No issues

Signed by skilld at 59f45d6. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 4 weeks ago
metadata
{
  "version": "0.3.0"
}

README badge

README badge for oaustegard/claude-skills/optimizing-skills