@cryptotaxi247 / SquadScope
Code Commits Issues Pulls Sessions CI/CD
main
main
md 4.37 KB
View source Raw

Nibbler's History

Core Context

  • Project: SquadScope (https://github.com/jmservera/SquadScope) — weekly editorial trend analysis, public site at https://jmservera.github.io/SquadScope/
  • User: jmservera
  • Created: 2026-05-25
  • Hired because: the team narrowly avoided shipping a "double S" logo that risked Nazi-SS-rune resemblance. The miss was caught by the human user, not by any agent. Nibbler exists so that class of failure-mode catch becomes the team's job, not the user's.

Why I'm here (in one sentence)

Catch what the makers miss. Before users do.

What's currently in flight or recently merged that I should sweep

  • Design proposal (PR #178 merged): docs/design/redesign-proposal-2026-05.md — color palette, tokens, layout direction
  • Icon spec (PR #178 merged, REPLACEMENT in flight via PR #186): docs/design/icon-spec.md — robot-with-binoculars replaces the rejected double-S design
  • Phase 1: tokens + typography (PR #181 merged): assets/css/tokens.css, font stacks
  • Phase 2: header + footer + nav (PR #185 merged): includes GitHub icon button in header
  • Cookie banner + privacy policy (in flight as #183, Hermes drafting): EU compliance, UX consent flow
  • GA4 analytics (in flight as #182): consent-gated tracking
  • Distribution strategy (in flight as #184, Morbo): announcement channel mix
  • Pipeline prompts: prompts/analyze-weekly.md, prompts/analyze-topic.md, prompts/reskill.md — the AI analysis loop

Team I work with

  • Hermes (Security) — code/secrets; I cover content/UX harms
  • Calculon (Design) — aesthetics; I sanity-check against hate-symbols
  • Amy (Frontend) — implementation; I review accessibility floor + dark-pattern surface
  • Farnsworth (Analyst) — editorial content; I review for bias, hallucination, harmful framing
  • Leela (Lead) — final approval; I'm a reviewer gate, not the final word

Learnings

2026-05-25 — First Sweep

References that paid off: - MS RAI Fairness + Transparency — gave immediate traction for the prompt injection finding (AI output transparency) and the GA4 fork-safety finding (privacy). These two principles map cleanly to concrete SquadScope artifacts. - OWASP LLM01 (Prompt Injection) — directly applicable to Farnsworth's {{RAW_JSON_CONTENT}} injection surface. The OWASP framing gave a precise vocabulary for the finding. - GDPR Articles 6(1)(a) + 5(1)(a) — Hermes had already done the heavy lifting on consent; my job was to verify the planned implementation didn't introduce dark patterns. The three-button consent modal passed all checks except one: button visual weight parity was not explicitly constrained. Low finding, but exactly the kind of implementation drift that turns a spec pass into a shipped dark pattern. - ADL Hate on Display — verified the current radar sweep icon is clean. The silhouette test (16px, rotation, inversion) revealed no ambiguity. The main lesson from the SS-icon incident: the failure mode hides at small sizes and under transformation. Always test at 16px first.

Artifact types and where they hide failure modes: - Icons/logos: Failure hides at 16px (favicon size) and under rotation/inversion. The craft-focused designer never tests this. - Consent UX: Failure hides in implementation drift from spec — spec says "equal prominence," implementation gives primary/ghost button treatment. Check CSS, not just copy. - Prompts: Failure hides in the injection surface — the description field from crawled repos is attacker-controlled. The instructions look safe in isolation; the danger is what happens when malicious repo descriptions reach the model. - Articles: Failure hides in AI authorship transparency (readers can't tell it's AI-generated) and in characterization claims (repo called "spam" without evidence citation). - Analytics: Failure hides in fork behavior — the GA4 tag silently reports to the repo owner's property from all forks.

Silhouette-test workflow developed: 1. Render SVG at 16px, 32px, 64px, 128px, 512px 2. Convert to grayscale 3. Rotate 90°, 180°, 270° 4. Invert colors (negative space check) 5. Mirror horizontally and vertically 6. Look for: letterforms (especially double letters), rune shapes, spoke patterns, hand gestures 7. Cross-reference ambiguous shapes against ADL Hate on Display database 8. Document findings in the icon spec as "Silhouette Safety Check ✓"