CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

tb-quarantine-triage

On-demand Tinybird quarantine investigator. Detects datasources with live quarantined rows, runs per-datasource root-cause analysis, presents a diagnosis plan for human review, then creates GitHub issues and git worktrees for approved datasources. No automation — always requires human sign-off before any GitHub issue or Git action.

Quality

83%

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

SKILL.md
Quality
Evals
Security

Quality

Content

85%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.

The body is an unusually actionable, well-sequenced seven-phase workflow with explicit human-approval gates and strong error handling — top marks for actionability and workflow clarity. The weaknesses are inlined detail that belongs in reference files and a referenced investigator prompt file that is missing from the bundle entirely, which caps progressive disclosure at 3.

Suggestions

Ship the referenced prompts/investigator.md file in the skill bundle (or inline its prompt) — Phase 2 depends on it, and it is currently absent, breaking the workflow at runtime.

Move the diagnosis-bundle JSON schema, the per-fix-type recovery playbook, and the issue body template into a reference file (e.g., references/diagnosis-format.md), keeping SKILL.md as a lean phase overview.

Trim rationale asides like "(materialized views, numbered MV shards, and copy targets — they do not receive direct ingestion)" to keep only the exclusion rules, since Claude does not need the why to execute the filter.

DimensionReasoningScore

Conciseness

The body is dense and operational with essentially no explanation of concepts Claude already knows — every line carries repo-specific facts (exclusion patterns, gh project field choreography, 1-month quarantine retention). It falls short of level 5 because the ~25-line recovery-path block inside the Phase 3 report template and the full 30-line diagnosis-bundle JSON schema are inlined where a tighter summary or reference file would do. It is clearly above level 3, which expects unnecessary explanation.

4 / 5

Actionability

Guidance is copy-paste executable throughout: exact SQL ("SELECT count() AS live_quarantined, max(insertion_date)"), exact exclusion regexes, gh issue/project commands with --json/--jq filters, a complete JSON issue payload, and a concrete git worktree fallback including the .git/info/exclude trick. Specific examples cover the common cases, matching the level-5 anchor; only placeholders like {DS_NAME} remain, which are parameterized by design.

5 / 5

Workflow Clarity

Seven phases are clearly sequenced with two explicit blocking validation checkpoints ("Do not proceed to Phase 5 until the user responds"; same before Phase 7), plus race-condition handling for null dominant_error, fingerprint-based duplicate-issue prevention, per-failure error handling, and an explicit stop condition when no live rows remain. This matches the level-5 anchor (explicit validation steps and feedback loops for error recovery); the approval gates validate the batch GitHub/worktree operations.

5 / 5

Progressive Disclosure

The body references ".claude/skills/tb-quarantine-triage/prompts/investigator.md", but no prompts/ directory or any other bundle file exists — a broken in-bundle reference. Additionally, content that clearly belongs in a separate file (the 30-line diagnosis JSON schema, the recovery-path playbook, the issue body template) is inlined in a 330-line SKILL.md. This matches the level-3 anchor (content that should be separate is inline; references not verifiable) rather than level 4, which expects references to be mostly clear and appropriately placed.

3 / 5

Total

17

/

20

Passed

Description

75%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 is specific, third-person, and highly distinctive, listing the full detect → diagnose → present → create-issues/worktrees pipeline. Its one real weakness is the absence of any explicit trigger guidance ("Use when..."), which caps completeness at 3, and slightly thin keyword synonyms.

Suggestions

Add an explicit trigger clause, e.g. "Use when the user mentions Tinybird quarantined rows, data quality errors in datasources, or asks to investigate quarantine/rejected rows."

Include a few natural synonyms users might say, such as "schema mismatch", "discarded/rejected rows", or "Tinybird data quality", to broaden trigger coverage.

Consider mentioning the diagnosis deliverable (a per-datasource root-cause report) more prominently as the entry-point use case, since that is the first user-visible outcome.

DimensionReasoningScore

Specificity

"Detects datasources with live quarantined rows, runs per-datasource root-cause analysis, presents a diagnosis plan for human review, then creates GitHub issues and git worktrees" lists multiple specific concrete actions covering the skill's full pipeline. This matches the anchor for comprehensive coverage rather than the level-4 anchor, which expects minor gaps in coverage.

5 / 5

Completeness

The "what" is clearly and concretely answered, but there is no "Use when..." clause or equivalent explicit trigger guidance — "On-demand" only weakly implies invocation context. Per the judging guidelines, a missing 'Use when' clause caps completeness at 3, which also matches the level-3 anchor (clear 'what', 'when' weakly implied).

3 / 5

Trigger Term Quality

Natural domain keywords are present — "Tinybird", "quarantine"/"quarantined rows", "root-cause analysis", "GitHub issues" — which a user facing this problem would plausibly say. It falls short of level 5 because there are no synonyms or common variations (e.g., "schema mismatch", "data quality", "discarded rows"), and it sits clearly above level 3's "missing common variations" since the core niche terms are all covered.

4 / 5

Distinctiveness Conflict Risk

"Tinybird quarantine investigator" with quarantine-table probing and human-sign-off issue/worktree creation occupies a clear niche with distinct triggers; overlap risk with generic git/GitHub skills is minimal. It matches the level-5 anchor rather than level 4's "minor overlap risk with closely related skills".

5 / 5

Total

17

/

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

allowed_tools_field

'allowed-tools' contains unusual tool name(s)

Warning

Total

15

/

16

Passed

Repository

linuxfoundation/crowd.dev

Path

.claude/skills/tb-quarantine-triage/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.