CtrlK
BlogDocsLog inGet started
Tessl Logo

evaluate-pr

Evaluate a PR the harness produced — walk the change, run the system together, and build a firm understanding before you merge it. Use after the harness hands a converged PR to you for review (the "Ready for your review" comment), or any time you want to deeply review an agent-authored PR. The human-attentive skill at the back of the chain; the mirror of /intent. Outcomes — merge, close, or fix-it-yourself-and-push - no handing work back to the loop.

61

Quality

71%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/human-loop/evaluate-pr/SKILL.md

The canonical home for this skill is evaluate-pr in tdg-ninja/context-specs-factory-ai

SKILL.md
Quality
Evals
Security

Quality

Content

60%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, actionable, and clearly sequenced evaluation workflow with real, well-signaled reference files. Its main weakness is verbosity — repeated restatements across the philosophy, steps, and hard-nevers inflate the body.

Suggestions

Consolidate the repeated 'no handing work back to the loop' point so it lives once (e.g. in E9 and the Hard nevers) rather than four times across the intro, E9, Step 7, and Hard nevers.

Split the dense E7 paragraph into a short rule plus a one-line rationale; the current block buries the key memory-authoring guidance under qualifications.

Move the E1-E9 philosophy block into a reference file and keep only the 2-3 load-bearing principles inline, tightening the body toward an overview-plus-steps shape.

DimensionReasoningScore

Conciseness

Noticeably verbose: the intro, the E1-E9 philosophy block, and the Hard nevers restate the same points multiple times (e.g. 'no handing work back to the loop' appears four times), and E7 is a single dense paragraph that could be trimmed.

2 / 5

Actionability

Provides concrete, executable commands (git fetch, git checkout --detach, gh pr diff, gh pr view --json, git push origin HEAD:feature/<f>, git checkout main) and specific files to read, with only minor deferral to the references for mechanics.

4 / 5

Workflow Clarity

Steps 0-8 are clearly sequenced with checkpoints (clean-tree preflight, understanding gate in Step 6, explicit human go-ahead before the destructive merge/close in Step 7); minor gaps are deferred to references.

4 / 5

Progressive Disclosure

Two one-level-deep references (references/walkthrough.md, references/verdict.md) are clearly signaled with one-line descriptions and are real bundle files; the body is well-sectioned, though the inlined E1-E9 philosophy is heavy enough to keep it from a 5.

4 / 5

Total

14

/

20

Passed

Description

83%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description: concrete actions and outcomes, explicit when-triggers, third-person voice, and a distinct niche. It edges toward harness-internal jargon that slightly limits natural trigger-term coverage.

DimensionReasoningScore

Specificity

Lists several concrete actions ('walk the change, run the system together') and concrete outcomes ('merge, close, or fix-it-yourself-and-push'), with only minor abstraction in 'build a firm understanding'.

4 / 5

Completeness

Explicitly answers both what ('Evaluate a PR...walk the change, run the system together') and when ('Use after the harness hands a converged PR to you for review...or any time you want to deeply review an agent-authored PR') with concrete trigger phrases.

5 / 5

Trigger Term Quality

Includes natural phrases a user would say ('deeply review an agent-authored PR', 'Ready for your review', 'review', 'merge'), though it leans on harness-internal jargon that a user would not always verbalize.

4 / 5

Distinctiveness Conflict Risk

Carves a clear niche as the human-attentive back-of-chain mirror of /intent with distinct triggers, with only minor overlap risk against generic PR-review skills.

4 / 5

Total

17

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
tdg-ninja/context-specs-claude-code
Reviewed

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.