agentsclimarketplace

Learn

Skill joshuaporth/appsec-skill/.cursor/skills/learn

Portable secure code review skill for AI coding agents — OWASP/CWE coverage, structured findings, and remediation guidance. Works with Cursor, Claude Code, Kiro, and Open Agent Skills.

Install
npx -y skills add joshuaporth/appsec-skill --skill learn

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Generates a post-run feedback learning brief from benchmark artifacts. Use after scoreboard/scoring outputs to extract recurring patterns, mistakes, wins, and next-step improvement actions for the benchmark process.

SKILL.md

1.9 KB, as published. Nobody here has run it

Learn

Create a separate feedback-learning output from benchmark artifacts. This skill is intentionally independent from scoreboard generation.

Purpose

Produce a concise learning brief that helps improve future benchmark runs. Focus on patterns and actions, not rerunning scoring.

Inputs (default)

  • benchmark/artifacts/scoring/*.txt (required)
  • benchmark/artifacts/findings/*.txt (optional but recommended)
  • generated scoreboard output (optional context)

If an input set is missing, continue with available artifacts and state limits.

Method

  1. Read all available scoring artifacts.
  2. Optionally correlate with findings artifacts for root-cause context.
  3. Extract repeated themes:
    • where outcomes were strong,
    • where outcomes were partial/incorrect/unknown,
    • recurring reasoning or evidence-quality issues,
    • recurring remediation-quality issues.
  4. Turn themes into concrete process improvements for the next run.

Output format

Use this structure:

# Learning brief

## Observed patterns
- <pattern 1>
- <pattern 2>
- <pattern 3>

## What improved outcomes
- <high-signal behaviors that correlated with correct judgments>

## What hurt outcomes
- <failure modes and recurring quality gaps>

## Next iteration actions
1. <specific action>
2. <specific action>
3. <specific action>

## Experiments for next run
- <small measurable process experiment>
- <small measurable process experiment>

Rules

  • Keep recommendations actionable and benchmark-specific.
  • Ground claims in artifact content; do not invent evidence.
  • Prefer 3-7 bullets per section and 3-5 actions total.
  • Keep this output separate from scoreboard output.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.