CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

post-mortem

Diagnose instruction defects and optionally submit Rosetta GitHub issue

Quality

79%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./instructions/r2/core/skills/post-mortem/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

96%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.

An exceptionally lean and executable instruction-only skill: a strictly ordered two-phase gated workflow with exact commands, verbatim templates, explicit verification checkpoints, and complementary principle/procedure/verification/failure-mode coverage with no padding. The only structural weakness is mild monolithicity — all ~120 lines inline with no reference split.

DimensionReasoningScore

Conciseness

The body is fully telegraphic ("Root cause, not symptoms: the defect is in the harness, not the artifact it produced") with zero pedagogical padding — it never explains concepts Claude already knows. Recurring themes (sanction gates, layer attribution) appear once each as principle, procedure, verification, and failure mode rather than as duplicated prose, so every token earns its place.

5 / 5

Actionability

Guidance is copy-paste ready: the exact command "gh issue create --repo griddynamics/rosetta --title <t> --body <b>", the exact sanction sentence "Submit the issue as drafted", a concrete fallback URL when gh is missing, and two complete fill-in templates. Not anchor 4 because there are no gaps in executable specifics.

5 / 5

Workflow Clarity

Sixteen strictly ordered steps across two hard-gated phases ("② NEVER starts without explicit opt-in"), with explicit checkpoints ("OUTPUT evidence inventory ... ONLY THEN proceed") and error-recovery feedback loops ("Comment/edit → revise → re-OUTPUT full draft → RE-ASK. Loop until sanctioned or declined") plus a dedicated validation checklist. This matches the top anchor exactly.

5 / 5

Progressive Disclosure

Well-organized single-file structure (core_concepts, process, validation_checklist, pitfalls, templates) with no bundle files and all content needed at every invocation inline. Not anchor 5 because at ~120 lines it exceeds the under-50-line well-organized single-file exception and offers no overview/navigation beyond the section tags; not anchor 3 because nothing that clearly belongs in a separate file is inlined — the templates are required on every run.

4 / 5

Total

19

/

20

Passed

Description

53%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.

The description states a clear, mostly distinct 'what' in two concrete actions but entirely lacks any 'when' trigger guidance and the natural trigger terms users would say. It sits at the midpoint of the rubric: clear on capability, weak on invocation cues.

Suggestions

Add an explicit trigger clause, e.g. "Use when a run fails or produces a defective artifact and the user asks what went wrong, wants a post-mortem, or wants to file the finding as a Rosetta issue."

Include the natural trigger terms users would actually say — "post-mortem", "what went wrong", "root cause analysis", "retrospective" — instead of only the technical phrase "diagnose instruction defects".

Enumerate the core actions for fuller coverage: attribute defects to harness layers (prompt / workspace files / config / instructions / tooling), produce a post-mortem report, and optionally file a sanitized GitHub issue.

DimensionReasoningScore

Specificity

The description names the domain and two concrete actions — "Diagnose instruction defects" and "optionally submit Rosetta GitHub issue" — matching the 1-2-action anchor. Not anchor 4 because coverage is incomplete (no mention of root-cause attribution, the post-mortem report, or layer analysis); not anchor 2 because two real actions are named rather than a generic domain label.

3 / 5

Completeness

A clear 'what' is present ("Diagnose instruction defects and optionally submit Rosetta GitHub issue") but there is no "Use when..." clause or equivalent trigger guidance, which caps completeness at 3 per the judging guidelines. Not anchor 4 because 'when' is entirely absent rather than merely imprecise.

3 / 5

Trigger Term Quality

Some relevant keywords are present ("diagnose", "instruction defects", "GitHub issue") but the natural phrases a user would actually say — "post-mortem", "what went wrong", "root cause", "retrospective" — are all missing. Not anchor 4 because most natural variants are absent rather than just a few; not anchor 2 because there are several domain-relevant keywords rather than one or two generic ones.

3 / 5

Distinctiveness Conflict Risk

The Rosetta-specific framing ("instruction defects", "Rosetta GitHub issue") carves a mostly distinct niche with only minor overlap risk against general review/diagnostic skills. Not anchor 5 because no trigger phrases exist and "diagnose defects" could be conflated with code-review-style skills; not anchor 3 because it is more specific than a broad "works with document files"-level overlap.

4 / 5

Total

13

/

20

Passed

Validation

93%

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

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

15

/

16

Passed

Repository

griddynamics/rosetta

Path

instructions/r2/core/skills/post-mortem/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.