agentsclimarketplace

Scoreboard

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

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 scoreboard

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

Builds the benchmark markdown scoreboard (table and Pulse) from `benchmark/artifacts/scoring/*.txt`. Use when the user asks for a scoreboard from current scoring artifacts without rerunning findings or scoring.

SKILL.md

2.7 KB, as published. Nobody here has run it

Scoreboard

Generate the scoreboard from existing scoring artifacts only.

Inputs

  • Read benchmark/artifacts/scoring/*.txt.
  • Each scoring artifact is expected to contain:
    • Challenge: NN
    • Primary intended issue: ...
    • Verdict: correct|partially_correct|incorrect
    • Rationale: ...
    • Bonus findings: ... (optional)
  • Do not read benchmark/challenges/ or rewrite findings during scoreboard generation.

Output format

Title block

# Benchmark scoreboard

━━━━━━━━━━━━━━━━━━━━━━  scored challenges · blind find -> judge  ━━━━━━━━━━━━━━━━━━━━━━

Glyphs

GlyphMeaning
🟢Hit (correct)
🟡Part (partially_correct)
🔴Miss (incorrect)
⚪Unknown (missing or malformed scoring artifact)

Results table

Use this header exactly:

|  #  |     | Intended issue (<=72 chars) | Judge notes (<=96 chars) |
| :-: | :-: | :--------------------------- | :----------------------- |

Rows:

  • Include discovered challenge rows only.
  • Sort rows by challenge id ascending.
  • One glyph per row (🟢/🟡/🔴/⚪).
  • Keep ASCII pipes and table column order unchanged.

Pulse

After the table, emit:

### Pulse

**🟢 n · 🟡 p · 🔴 m** - <one sentence>

Replace n, p, m with counts from the rendered table. If any unknown rows exist, mention the unknown count (⚪) in the sentence.

Parsing and row construction

  1. Scan benchmark/artifacts/scoring/*.txt.
  2. For each file, extract NN from filename and/or Challenge:.
  3. Determine row status:
    • 🟢 for Verdict: correct
    • 🟡 for Verdict: partially_correct
    • 🔴 for Verdict: incorrect
    • ⚪ if required fields are missing/malformed
  4. Build columns:
    • #: two-digit NN
    • Intended issue: Primary intended issue (truncate to 72 chars)
    • Judge notes: Rationale (optionally enrich with Bonus findings, non-redundant only, truncate to 96 chars)
  5. Unknown-row fallback text:
    • Intended issue: - (unknown scoring artifact)
    • Judge notes: Unknown scoring output for this challenge.
  6. If two files map to the same NN, keep the newest file and mention deduplication in the pulse sentence.

What not to do

  • Do not output unsorted or duplicate challenge rows.
  • Do not add/remove/reorder table columns.
  • Do not change glyph meanings.

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.