CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

lightrun-runtime-aware-pr-review

Use when reviewing a pull request with runtime or production evidence — for example to review a PR with runtime verification, gather production evidence, or simulate a patch on live samples. Reviews a pull request by diffing against the PR merge base, collecting live samples, and simulating the patch on captured production inputs.

1.38x
Quality

90%

Does it follow best practices?

Impact

98%

1.38x

2 of 3 eval scenarios. Add 1 more for a full score.

SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

81%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 orchestration body: clear goal/scope, ordered phase references, explicit exit-criteria checkpoints with feedback loops, and hard working rules. The main issues are verbatim duplication of constraints.md in Working Rules and four dead, conflicting reference files left in the bundle.

Suggestions

Delete the four stale unreferenced files in references/ (phase-2-runtime-profile.md, phase-3-sampling-plan.md, phase-4-execute-snapshots.md, phase-5-patch-simulation.md) from the old phase numbering — their content conflicts with the current phase files and misleads when the bundle is listed.

Trim Working Rules bullets that restate constraints.md verbatim (the PR-diff rule, the unmapped-coordinates rule, the async-snapshot rule) down to the pointer 'Follow every constraint in Constraints' plus only phase-flow-specific rules.

Add one minimal inline example of a phase artifact or stop condition (e.g. the shape of a snapshot entry or the Phase 1 not-verifiable message) so the checkpoints' criteria are unambiguous without opening a reference file.

DimensionReasoningScore

Conciseness

The body is lean with no explanations of concepts Claude already knows, but the Working Rules section restates several constraints nearly verbatim ("Never use `deployed_sha` → `pr_head_sha` as the PR review diff"; "Never query, instrument, or snapshot PR-only or unmapped coordinates") that references/constraints.md already owns, despite also pointing to it. This is the 'minor instances that could be trimmed' anchor rather than the fully lean anchor.

4 / 5

Actionability

Guidance is concrete and specific for an orchestration body — ordered reference reading, per-phase artifacts ("snapshot entries exist for the highest-value uncovered verification areas (up to 3)"), and identifier-level rules — but all executable detail (commands, message formats, examples of a Phase 1 stop or Sampling Request) is delegated to reference files with no inline example. This fits 'mostly executable guidance with minor gaps' rather than fully executable.

4 / 5

Workflow Clarity

The phases are clearly sequenced (0–6 in order) with an explicit per-phase exit-criteria checklist and genuine feedback loops: "Do not advance a phase until its exit criteria are satisfied; loop back to the prior phase when evidence is incomplete" and the Phase 5 Sampling Request stop protocol for zero-hit locations. This matches the anchor 'Clear sequence with explicit validation steps; feedback loops for error recovery; checklists for complex processes'.

5 / 5

Progressive Disclosure

The body is a clear overview with a well-signaled, numbered, one-level-deep reference list and all 10 referenced files exist. However, the actual bundle contains four stale unreferenced files from an older phase numbering (phase-2-runtime-profile.md, phase-3-sampling-plan.md, phase-4-execute-snapshots.md, phase-5-patch-simulation.md) whose content conflicts with the current phase files (e.g. the old runtime-profile check vs. the current 'coming soon' stub), creating a navigation hazard when the bundle is listed. This fits 'good structure with minor organization gaps' rather than the clean anchor 5.

4 / 5

Total

17

/

20

Passed

Description

96%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: concise, concrete, third-person, with an explicit 'Use when' clause and specific capability verbs covering the full pipeline. The only weakness is that its lead trigger phrase overlaps with generic PR-review skills, slightly raising conflict risk.

DimensionReasoningScore

Specificity

"Reviews a pull request by diffing against the PR merge base, collecting live samples, and simulating the patch on captured production inputs" names three concrete actions that comprehensively cover the skill's pipeline. It fits the anchor 'Lists multiple specific concrete actions; comprehensive coverage' rather than 4, since no meaningful action in the workflow is left out.

5 / 5

Completeness

It explicitly answers when ("Use when reviewing a pull request with runtime or production evidence — for example to review a PR with runtime verification, gather production evidence, or simulate a patch on live samples") and what ("Reviews a pull request by diffing against the PR merge base, collecting live samples, and simulating the patch"). Both halves are explicit with concrete trigger phrases, matching the anchor-5 example structure.

5 / 5

Trigger Term Quality

Natural trigger terms include "reviewing a pull request", "PR", "runtime verification", "production evidence", "live samples", "patch", and "merge base", with synonyms covered (pull request/PR, runtime/production). This matches the comprehensive-coverage anchor; a user needing this skill would naturally use these phrases.

5 / 5

Distinctiveness Conflict Risk

The runtime-evidence niche is clearly distinguishable from generic skills, but the lead phrase "reviewing a pull request" is a common trigger for ordinary code-review skills, creating minor overlap risk unless the "with runtime or production evidence" qualifier is honored. This fits 'Mostly distinct; minor overlap risk with closely related skills' rather than 5.

4 / 5

Total

19

/

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.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

lightrun-platform/lightrun-ai

Path

skills/lightrun-runtime-aware-pr-review/SKILL.md

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.