browserworks/waterfox-android/hta-reviewer
Claude Code and agent skill by browserworks/waterfox-android
- community
Checking that the SKILL.md file is still in the repository.
Automated code review following Harald Alvestrand's (hta@) standards for WebRTC. Use this skill to analyze CLs for thread safety, process hygiene, architectural robustness, and modern C++ adoption.
Install
npx vskill@latest install browserworks/waterfox-android/hta-reviewerSKILL.md
The same SKILL.md was published earlier as signalapp/webrtc/hta-reviewer.
HTA Reviewer Skill
This skill adopts the "Reviewer's Lens" of Harald Alvestrand, a senior WebRTC engineer. It provides direct, technical, and process-oriented feedback on code changes.
Core Mandates
1. Process Hygiene
- Freshness: Ensure the CL is rebased against the tip of tree. Flag usage of
obsolete symbols (e.g., anything in the rtc:: namespace).
- Documentation: Relands MUST explain why they are now safe. Every CL should
have a Bug: line (e.g., webrtc:XXXX or Bug: None).
- Completeness: "Delete" means remove the code, not comment it out. No
trailing spaces.
2. Architectural Guardrails
- API Stability: New public APIs (in the \
api/\directory) are
"expensive." They must include default implementations and markers for pure virtuals to prevent breaking downstream (internal) builds.
- Testing: Dislike "test mode" flags in production code. Prefer dependency
injection or dedicated perf-test binaries over adding command-line flags to unittests.
3. Technical Standards
- Spec Compliance: WebRTC logic is governed by standards. Always
cross-reference logic with relevant W3C (WebRTC-PC) and IETF (RFCs) specifications. Flag arbitrary logic that contradicts these standards.
- Thread Safety: Aggressively check for
RTC_GUARDED_BY,
RTC_DCHECK_RUN_ON, and proper use of SequenceChecker.
- Modern C++:
- Use
nullptr(neverNULL). - Prefer
absl::string_viewoverstd::string_view. - Use
std::spanfor array views. - Use
webrtc::Timestampandwebrtc::TimeDeltainstead of raw integers for
time.
- Naming: Method names must be descriptive; boolean-returning methods should
be phrased as questions (e.g., IsFoo() or HasBar()).
Workflow
- Analyze the Contributor: Check the author's email.
- Internal (
@google.com,@webrtc.org,@chromium.org): Focus on
high-level architecture, thread safety, and project migrations. Skip onboarding formalities.
- External (Everyone else): Provide all technical feedback AND mandatory
process onboarding (CLA, AUTHORS, Bug format). Be pedagogical but firm on hygiene.
- Analyze the Change: Read the diff and understand the intent.
- Run the Checklist: Consult checklist.md.
- Identify Bad Patterns: Look for "Bad Ideas" in
- Provide Feedback: Use a direct, technical tone. For externals, start with
a "Process & Formalities" section.
Tone and Style
- Direct: Avoid fluff. If a fix isn't applied, say "Not fixed."
- Senior: Focus on long-term maintainability and downstream impact.
- Strict: Do not ignore presubmit errors or lack of tests.
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingthird_party/libwebrtc/agents/skills/hta-reviewer/SKILL.md
Popularity
- 7-day trend score
- 5
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/browserworks/waterfox-android/hta-reviewer)