CtrlK
BlogDocsLog inGet started
Tessl Logo

fusion-github-review-resolution

Resolves unresolved GitHub PR review threads end-to-end: evaluates whether each review comment is correct, applies a targeted fix when valid, replies with rationale when not, commits, and resolves the thread. USE FOR: unresolved review threads, PR review feedback, changes requested PRs, PR review URLs (#pullrequestreview-...), fix the review comments, close the open threads, address PR feedback. DO NOT USE FOR: summarizing feedback without code changes, creating new PRs, or read-only branches.

85

1.06x
Quality

81%

Does it follow best practices?

Impact

93%

1.06x

Average score across 3 eval scenarios

SecuritybySnyk

Advisory

Suggest reviewing before use

SKILL.md
Quality
Evals
Security

Evaluation results

92%

16%

PR Review Resolution Workflow

End-to-end review resolution workflow

Criteria
Without context
With context

Checklist document created

100%

100%

Phase order documented

80%

100%

Reply before resolve

100%

100%

Specific reply content

100%

100%

Baseline captured

75%

100%

Single push described

50%

100%

Outdated thread handled correctly

100%

100%

Incorrect feedback identified

100%

100%

No secrets in replies

100%

100%

One commit per comment

50%

100%

Re-review question noted

0%

50%

Worktree question noted

0%

0%

100%

Automating Review Thread Resolution with Shell Scripts

Scripted tooling and dry-run workflow

Criteria
Without context
With context

Dry-run command first

100%

100%

Correct apply flag usage

100%

100%

Correct PR arguments

100%

100%

Duplicate reply guard explained

100%

100%

allow-additional-reply condition

100%

100%

MCP tools preferred first

100%

100%

Bundled assets as second choice

100%

100%

Ad hoc scripts prohibited

100%

100%

get-review-comments script referenced

100%

100%

Dry-run default explained

100%

100%

87%

High-Volume PR Review with GraphQL Resolution

Comment judgment and GraphQL rate-limit strategy

Criteria
Without context
With context

Incorrect feedback identified

100%

100%

Evidence-based decline reply

100%

100%

Read-through cache decline

100%

100%

resolveReviewThread uses threadId

100%

100%

Reply mutation variable names correct

70%

100%

Rate-limit threshold warning

85%

71%

Mutation pacing documented

100%

100%

Reply before resolve per thread

100%

100%

Total mutation count estimated

62%

50%

Fetch once, reuse

50%

50%

Repository
equinor/fusion-framework
Evaluated
Agent
Claude Code
Model
Claude Sonnet 4.6

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.