CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

data-research

Structured data research: search sources, extract structured data, archive raw sources, maintain canonical tracker pages, deduplicate. Parameterized via YAML recipes for investor updates, donations, company updates, or any email-to-structured-data pipeline.

Quality

82%

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

88%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 lean, well-sequenced 7-phase pipeline with strong validation culture (integrity rule, dedup checkpoint, explicit error recovery) that keeps every section operational. The main gaps are minor: a few vague pointers ('enrichment service', 'Brain first'), no example recipe YAML inline, and inconsistent reference path notation.

DimensionReasoningScore

Conciseness

The body is dense and operational throughout — specific paths ('~/.gbrain/recipes/{name}.yaml'), tool names, dedup tolerances, and a concrete output table — with zero explanation of concepts Claude already knows and no padded sections. The two-line intro after the H1 mirrors the frontmatter description, but that is standard standalone structure rather than over-explanation, so this matches 'Lean and efficient; every token earns its place' rather than 4.

5 / 5

Actionability

Guidance is concrete and executable for an instruction-only skill: exact tool names (put_raw_data, file_upload, gbrain write-request <request_id>), exact recipe paths, deterministic dedup rules with tolerances, and a copy-ready output markdown table. It is not 5 because a few steps stay high-level — 'Brain first', 'Uses enrichment service for entity pages', and the recipe YAML field structure is never shown, leaving minor gaps.

4 / 5

Workflow Clarity

Seven clearly sequenced phases with explicit validation for the batch operations: the numbered EXTRACTION INTEGRITY RULE ('re-read from saved files', 'Never trust LLM working memory'), a dedup checkpoint before tracker writes with exact/fuzzy/conflict handling, a fail-improve loop logging LLM fallbacks, and a dedicated error-recovery section with per-error recovery steps. This matches 'Clear sequence with explicit validation steps; feedback loops for error recovery' — the batch-operation cap does not apply since validation is present.

5 / 5

Progressive Disclosure

The single-file body (~120 lines) is well organized into clearly signaled sections, and external references (recipes directory, agent-operator-protocol.md, conventions/quality.md) are one level deep and explicitly pathed. It is not 5 because no bundle reference files exist at all — recipe format details and the conventions material could be split out — and path notation is inconsistent ('../conventions/agent-operator-protocol.md' vs 'skills/conventions/quality.md'). It is not 3 since nothing inlined clearly belongs in a separate file.

4 / 5

Total

18

/

20

Passed

Description

71%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, specific, third-person description that clearly names the pipeline's actions and its recipe-based domains, dragged down by the absence of an explicit use-when clause and incomplete natural-term coverage. Adding one 'Use when...' sentence covering expenses/receipts-style triggers would lift both completeness and trigger term quality.

Suggestions

Append an explicit trigger clause, e.g.: 'Use when the user wants to track investor updates, donations, expenses, or company metrics, or says research, track, or build a tracker.'

Include the expense/receipt domain from the body's built-in recipes (e.g., 'expenses', 'receipts', 'recurring charges') so users phrasing requests around spending also match.

Add natural synonyms users would say such as 'log' or 'keep a tracker of' alongside the current domain keywords.

DimensionReasoningScore

Specificity

The description lists five concrete, distinct actions — "search sources, extract structured data, archive raw sources, maintain canonical tracker pages, deduplicate" — which is comprehensive coverage of the pipeline in third-person voice. It matches the anchor 'Lists multiple specific concrete actions; comprehensive coverage' and is not a 4 since no meaningful action in the pipeline's scope is left unmentioned.

5 / 5

Completeness

The 'what' is clearly stated (search, extract, archive, maintain trackers, deduplicate), but there is no explicit 'Use when...' clause — the phrase "Parameterized via YAML recipes for investor updates, donations..." only weakly implies when to invoke the skill. Per the judging guideline that a missing explicit trigger guidance caps completeness at 3, and matching the anchor 'Has a clear what but when is missing or only weakly implied', this is a 3, not a 4, since the when-guidance is inferred rather than stated.

3 / 5

Trigger Term Quality

Natural user phrases like "investor updates", "donations", "company updates", and "structured data" are present, but common variations a user would actually say — "expenses", "receipts", "track spending", "log" — are absent (the body's expense-tracker recipe never surfaces in the description). This fits 'Good keyword coverage; a few natural terms missing'; it is not 5 because synonym coverage is incomplete, and not 3 because the included terms are genuinely natural rather than technical jargon.

4 / 5

Distinctiveness Conflict Risk

The email-to-structured-data framing plus named domains (investor updates, donations, company updates) carves a fairly distinct niche, but generic leading terms like "research" and "search sources" create minor overlap risk with general research or email-processing skills. This fits 'Mostly distinct; minor overlap risk with closely related skills' rather than 5, where the triggers themselves would be uniquely identifying.

4 / 5

Total

16

/

20

Passed

Validation

87%

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

Validation — 14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

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

Warning

relative_links

Relative link issues: 1 missing, 1 suspicious

Warning

Total

14

/

16

Passed

Repository

garrytan/gbrain

Path

skills/data-research/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.