41 Commits
Author SHA1 Message Date
Wayner Barrios 66eac6e8db feat: add Claude Code packaging and skill quality gates (#26)
- add native Claude Code plugin and marketplace manifests
- document Claude Code installation, invocation, updates, and removal
- reduce routing metadata and enforce a catalog budget
- validate Agent Skills and Claude manifests in CI
- remove wooyun-legacy from the package and provenance
- update package metadata and tests for the 54-skill catalog
2026-08-13 17:06:24 -04:00
Wayner Barrios 086e9bae2f Add selective npm skill installer and distinguish code review workflows
Adds selective skill installation, curated profiles, dependency expansion,
safe replacement, provenance preservation, distinct comprehensive and focused
code-review identities, complete documentation, and npm release preparation
for version 0.4.0.
2026-08-13 16:25:44 -04:00
Wayner Barrios ab04919087 Add 10 more piolium security skills (MIT)
GitHub Actions prompt-injection audit, threat modeling, differential
review, variant analysis, SARIF parsing, Semgrep/CodeQL scanning and
rule authoring, and WooYun-derived web vuln methodology. Generalized
semgrep's piolium-specific subagent references. Skill count 45 -> 55.
2026-08-02 15:54:18 -04:00
Wayner Barrios f2695df75b Add 25 Hugging Face skills (Apache-2.0)
Ports huggingface/skills: hf CLI, cloud/SageMaker, training (TRL,
vision, sentence-transformers), Spaces/Gradio, papers, and
Transformers.js. Condensed 4 oversized skills, full UPSTREAMS.json
provenance, skill count 20 -> 45.
2026-08-02 14:25:44 -04:00
Wayner Barrios fbad717a23 Add 3 firstpick + 5 piolium skills (MIT)
Code quality, design patterns, paper summarization, plus 5 security
review skills (supply-chain risk, sharp edges, insecure defaults,
false-positive check, vuln reporting). Full UPSTREAMS.json provenance,
skill count 11 -> 19.
2026-08-02 14:01:03 -04:00
Wayner Barrios dd64d4a7e0 Add ai-slop skill, ported from omp-designer (MIT)
Adds a design rubric for detecting generic AI output, with full
UPSTREAMS.json provenance. Updates skill count from 11 to 12 across
package metadata, README, and tests.
2026-08-02 13:34:10 -04:00
Wayner Barrios 7f0ab9210d Add Pi package support and installation 2026-07-31 19:49:44 -04:00
Wayner Barrios 56620910be Highlight Codex support and installation 2026-07-31 19:38:01 -04:00
Wayner Barrios 9a41e8621a Add Codex plugin distribution 2026-07-31 19:22:41 -04:00
Wayner Barrios 81a7048aa6 Address whole-branch review findings
- write reports and snapshots atomically, cleaning temp files on failure
- isolate CLI run tests from repository artifacts via env overrides
- cover version-detection failures, unsupported modes, manifest bounds,
  redaction and grading edges, and malformed mutations with tests
- document path overrides and completed plan steps
2026-07-31 17:28:13 -04:00
Wayner Barrios d7bdd12729 Record reviewed behavioral baseline 2026-07-31 16:51:03 -04:00
Wayner Barrios d2e9b4ee88 Document behavioral evaluation workflow 2026-07-31 16:50:05 -04:00
Wayner Barrios b0a4780acc Add snapshot acceptance and replay to behavioral evaluations 2026-07-31 16:50:05 -04:00
Wayner Barrios cd5c150b08 Harden behavioral evaluation isolation 2026-07-31 16:50:05 -04:00
Wayner Barrios 4a74b6f013 Add opt-in behavioral evaluation runner 2026-07-31 16:50:05 -04:00
Wayner Barrios 6e6b7dd698 Define adversarial evaluation corpus 2026-07-31 16:50:05 -04:00
Wayner Barrios 8feef8cefa Add deterministic behavioral evaluator 2026-07-31 16:50:05 -04:00
Wayner Barrios e3a1a098e7 Clarify Claude rule loading 2026-07-28 17:06:23 -04:00
Wayner Barrios b63665efd8 Enforce prompt safety recovery 2026-07-28 16:11:46 -04:00
Wayner Barrios af67900ca3 Harden prompt trust boundaries 2026-07-28 15:57:23 -04:00
Wayner Barrios f2e826d874 Fix portable rule resource loading 2026-07-28 15:48:10 -04:00
Wayner Barrios 15a497f766 Correct project rules resolution 2026-07-28 15:38:42 -04:00
Wayner Barrios 95c6ffbedd Clarify MCP negotiation tests 2026-07-28 15:30:21 -04:00
Wayner Barrios 35257451b1 Modernize MCP server guidance 2026-07-28 15:18:57 -04:00
Wayner Barrios 7ccfa329af Separate frontend planning scopes 2026-07-28 14:09:20 -04:00
Wayner Barrios ad211a879c Complete review scope guarantees 2026-07-28 11:41:42 -04:00
Wayner Barrios ec49e95cbe Freeze security review scope 2026-07-28 11:30:37 -04:00
Wayner Barrios 8d4eacbba6 Harden security review contracts 2026-07-27 15:33:27 -04:00
Wayner Barrios ac297607ff Clarify code review recovery order 2026-07-27 15:27:14 -04:00
Wayner Barrios 8d80e72bde Harden code review orchestration 2026-07-27 15:21:30 -04:00
Wayner Barrios 5f6f2e2531 Preserve incomplete workflow coverage 2026-07-27 15:08:31 -04:00
Wayner Barrios a3714c09a5 Define feature workflow handoffs 2026-07-27 15:04:30 -04:00
Wayner Barrios c753cd35ad Adopt native skill discovery 2026-07-27 14:32:22 -04:00
Wayner Barrios 6e83761707 Register read-only workflow agents 2026-07-27 14:32:04 -04:00
Wayner Barrios 559c97219d Add immutable skill provenance 2026-07-27 13:25:30 -04:00
Wayner Barrios 318ca65626 Add working-discipline guidance to the code and authoring skills
feature-dev, code-architect, and mcp-builder gain a shared "Working
discipline" section (think before acting, simplicity first, surgical changes,
goal-driven). code-review and code-reviewer get a lighter touch — a "scope
creep" criterion added to their review checklists. security-review and
skill-creator already embody this discipline, so they are not modified, and
frontend-design is left alone because "simplicity first" would clash with the
skill's deliberate bold/maximalist direction.

The five command files are regenerated from their skills (in sync). A new
tests/working-discipline.test.mjs guards the additions so the rule cannot
silently regress.
2026-05-28 12:37:42 -04:00
Wayner Barrios a2b5459e35 Make agents-md-revise respect first-match-wins rules loading
The skill told the model to "edit whichever rules file the project already
uses." Under OpenCode's first-match-wins resolution, a CLAUDE.md co-located
with an AGENTS.md is never read, so learnings captured there would be silently
lost. The skill now describes first-match-wins and directs capture to
AGENTS.md when both exist (never the shadowed CLAUDE.md), reinforced in the
"where each addition belongs" step. A test guards it; command kept in sync.
2026-05-27 22:53:03 -04:00
Wayner Barrios 3aa3153d53 Enforce variety across generations in frontend-design
The skill already warned against converging on common choices, but the new
font and layout lists risk becoming the next default the way Space Grotesk
did. This adds an explicit "vary across generations" rule — treat the lists
as seeds to rotate, deliberately diverge when a brief resembles a prior one,
and never ship two different products as the same template. A matching
self-critique check rejects output that reads as a default pattern or a
previous generation. A test guards the rule; the command is regenerated from
the skill.
2026-05-27 15:53:52 -04:00
Wayner Barrios 09dc86739a Add design specifics and a self-critique rubric to frontend-design
The skill was pure aesthetic prose with no self-review. This adds three
sections, reworded from cited sources and kept under 500 lines:

- Concrete anti-default moves: font-by-vibe picks, layout archetypes, type
  contrast, semantic CSS-variable tokens, one orchestrated page-load.
- An objective self-critique rubric the model runs before returning (WCAG 2.2
  AA contrast/semantics/alt/focus/reflow plus craft checks), framed honestly
  since models tend to over-praise their own output.
- An optional, cross-platform validation note (html-validate static; pa11y or
  axe for rendered DOM), including an opt-in vision-only Playwright screenshot
  loop that text-only models skip.

Adds tests/frontend-design.test.mjs to guard the additions. The command is
regenerated from the skill and stays in sync.
2026-05-27 15:53:52 -04:00
Wayner Barrios f9bdd3ff74 Remove duplicated HIGH-SIGNAL block in code-review, add a dup test
The "Critical: only HIGH-SIGNAL issues should be flagged" block appeared
twice, back to back, in code-review's SKILL.md and command. The second copy
was a strict subset of the first (it dropped the "Concrete edge cases"
bullet), so it was pure redundancy. The duplicate is removed; the complete
first copy stays.

The skill-lint suite gains a per-skill "no duplicated block" check (no
non-empty line >= 50 chars repeats within a SKILL.md or its command), which
catches this class of copy-paste error. SKILL.md and command stay in sync.
2026-05-26 23:35:53 -04:00
Wayner Barrios 2dc168655f Add skill/command lint and restore truncated command descriptions
A generation step had truncated nine command descriptions at ~200
characters, cutting them mid-sentence. Each commands/*.md description is
restored to match its SKILL.md source, so the slash-command menu shows the
full text and skill and command stay in sync.

Adds a zero-dependency node:test suite (tests/skill-lint.test.mjs) that
enforces the invariants going forward: directory name matches the
frontmatter name and slug regex, description length within 1..1024 and not
truncated, SKILL.md body under 500 lines, and each command inlines its
SKILL.md body, shares its description, and ends with $ARGUMENTS. A GitHub
Actions workflow runs the suite on every push and pull request.
2026-05-26 23:09:45 -04:00