Hugging App 0.1.6

Symlink-safe corpus reads with regression tests; smaller icon. Bundle built from e105d57.

Hugging App 0.1.5

Approved happy phone-hugging mascot, native iPhone app guidance, and evidence-based design review. Bundle SHA-256: b7bdea0df593ea6db655dd46ef5370d42917f3441267977da161b3e3b19588d5.

App Design Research v0.1.1

Changelog

0.1.1

• Distinguish historical, synthetic, source-only and visually inspected evidence.
• Return up to three findings with validation steps and reusable review receipts.
• Add setup instructions, FAQ and a synthetic text-review example.
• Include required directory icon references and versioned deterministic bundles.
• Keep network research disabled; the separate Hermes compatibility source remains unchanged.

0.1.0

• Initial supplied-reference design review plugin.

App Design Research v0.1.0

App Design Research v0.1.0 provides an agent skill for reviewing screenshots, text, and code supplied by the user. It preserves provenance and separates observed evidence from proposed improvements.

The included Hermes compatibility adapter returns disabled responses before website dispatch. This release does not fetch AppLlama content, provide premium access, or include third-party screenshots or skill bundles.

Install in Codex

codex plugin marketplace add PatchworkMD/app-design-research --ref v0.1.0
codex plugin add app-design-research@app-design-research-public

Start a new task after installation.

Verification

• Five behavior tests passed, including all operations blocked before dispatch and no request-flag override.
• Plugin and skill validators passed.
• Version-tag installation succeeded; installed manifest, skill and icon matched the release source byte for byte.
• The skills ZIP uses a seven-file allowlist. SHA256SUMS is attached.

Read the privacy policy, terms and data acknowledgements, and sources.

ChatGPT directory submission is prepared but not submitted or approved. Verified publisher identity and OpenAI review remain required. Model-level evaluation cases are documented as expected behavior, not passed evaluations. No legal clearance is claimed.