Slang Review Resolve Judgment Calls
Testing & Qualityslang-review-resolve-judgment-callsSlang Review Resolve Judgment Calls is an Agent Skill for testing & quality published in the shader-slang/slang repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
Resolve Slang clarity review candidates marked as needing a judgment call. Use after consolidation and scope filtering, before posting, to perform focused follow-up analysis and decide whether uncertain candidates should be kept, revised, or dropped.
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Inspect the Agent Skill at https://github.com/shader-slang/slang/blob/HEAD/.claude/skills/slang-review-resolve-judgment-calls/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/slang-review-resolve-judgment-calls, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
Slang Review Judgment-Call Resolution
Review candidates marked Status: Needs judgment call or with Overlap decision: Needs judgment call. The goal is to resolve uncertainty through focused analysis, not to wait for a
human reviewer.
Do not post comments. Update the canonical candidate markdown file in place.
Inputs
Read:
- The canonical candidate markdown file after consolidation and scope filtering.
tmp/pr-diff.patchandtmp/pr-files.txt, or the equivalent PR diff and file list.- PR-version and base-version source around the candidate location.
- Callers, callees, tests, comments, and related declarations needed to answer the specific uncertainty recorded in the candidate.
Decisions
For each judgment-call candidate, spend focused effort to make a decision:
Status: Keepwhen the concern is credible, in scope, and useful to post.Status: Revisewhen the concern is credible but the proposed comment needs a narrower location, clearer wording, or a different request.Status: Dropwhen the concern is unsupported, duplicate, superseded, out of scope, or too speculative after follow-up analysis.Status: Needs judgment callonly when focused analysis still cannot resolve the question.
Update or add:
- Judgment decision: Keep | Revise | Drop | Needs judgment call
- Judgment rationale: <one or two sentences describing the decisive evidence or remaining uncertainty>
If keeping, revising, or leaving a candidate as Needs judgment call, make sure it still has
postable metadata:
- Scope decision: Direct | Contextual
- Scope rationale: <one sentence>
- Overlap decision: Keep
- Overlap rationale: <one sentence>
Working Process
- Inventory all candidates marked
Needs judgment call. - For each one, read the recorded uncertainty and identify the smallest source context needed to answer it.
- Follow local call graph, data-flow, type, declaration-reference, test, or documentation evidence far enough to decide whether the concern is credible and useful.
- Revise the proposed comment when the concern is real but the wording overstates confidence or points at the wrong location.
- Move dropped candidates to a later
## Droppedsection when practical, preserving enough rationale for auditability.