concept-synthesis

Deduplicate and synthesize raw concept stubs into a tiered intellectual map (T1 Canon to T4 Riff), tracing idea evolution across sources over time. Transforms thousands of raw concept pages into a curated intellectual fingerprint. Includes a reversible curation cull pass (Phase 5) with hard keep/delete/merge verdicts, substance gates, grounding labels, cluster budgets, and merge-with-backlinks salience promotion.

Quality

64%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./skills/concept-synthesis/SKILL.md

Quality

Content

70%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-sequenced, heavily validated multi-phase pipeline with genuinely strong workflow clarity and mostly concrete guidance. Its weaknesses are repetition-driven verbosity and a monolithic structure that inlines a large self-contained rubric instead of splitting it into a reference file.

Suggestions

Move the Phase 5 cull rubric (6-axis tables, verdict thresholds, grounding labels, cluster budgets) into a references/cull-rubric.md and keep a short summary plus a one-level-deep link in SKILL.md.

Cut repeated statements of the same idea (reversibility, test-before-bulk, distinctness guard) to a single authoritative statement each, and drop the 'Contract'/'Output Format' sections that only exist 'for the conformance test'.

Give Phase 2 a concrete tier-assignment rule (formula or thresholds) so the scoring step is as executable as the Phase 5 rubric.

Dimension	Reasoning	Score
Conciseness	The Phase 5 rubric tables and quality gates are domain-specific content Claude cannot already know, but the body repeats itself — reversibility is restated ~4 times ('fully reversible', 'nothing is destroyed', 'keep every undo path alive'), test-before-bulk appears 3 times, the distinctness guard is explained twice, and the 'Contract'/'Output Format' sections are boilerplate that admit they exist 'for the conformance test'.	3 / 5
Actionability	Concrete gbrain commands (query, search, put, delete, restore, history) and verbatim output templates make Phases 1 and 5 executable, and the merge mechanics are copy-paste ready. Not 5: Phases 2-4 remain high-level ('agent scores each canonical concept... and writes tier into frontmatter') with no tier thresholds, and no scoring formula beyond the Phase 5 rubric.	4 / 5
Workflow Clarity	Five phases are clearly sequenced with explicit validation for destructive batch operations: test-before-bulk spot-checks, a merge-quality gate that rejects incomplete merges and flags clusters for review, soft-delete plus _merged/ archive plus version history as undo paths, and a Quality gates + Anti-Patterns checklist. Matches the top anchor: feedback loops for error recovery and checklists for a complex process.	5 / 5
Progressive Disclosure	No bundle files exist (references/, scripts/, assets/ are absent) and the skill is a single ~500-line monolith in which the ~200-line Phase 5 cull rubric clearly belongs in a separate reference file. Internal section structure and external convention links (quality.md, test-before-bulk.md) are well-signaled, which keeps it above 'minimal structure', but the file is detail rather than an overview.	3 / 5
	Total	15 / 20 Passed

Description

58%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 specific, well-niched description with a strong 'what' but no 'when' — it never tells Claude when to reach for the skill, and the trigger phrasing users would naturally say is absent from the description itself. Third-person voice is correct throughout.

Suggestions

Append an explicit trigger clause, e.g. 'Use when the user asks to dedupe, synthesize, or cull their concepts, build their intellectual map, or decide which concepts to keep.'

Add natural user phrasings and synonyms ('dedupe my notes', 'concept cleanup', 'which ideas keep recurring') to the description itself rather than leaving them in the frontmatter triggers list.

Replace the jargon tail 'merge-with-backlinks salience promotion' with a plain action ('merges duplicates into canonical pages while preserving sources and angles').

Dimension	Reasoning	Score
Specificity	Lists several concrete actions — 'Deduplicate and synthesize raw concept stubs', 'tracing idea evolution across sources over time', 'hard keep/delete/merge verdicts, substance gates, grounding labels, cluster budgets' — with minor coverage gaps. Not 5: the tail 'merge-with-backlinks salience promotion' is jargon that gestures at a capability rather than naming an action.	4 / 5
Completeness	The 'what' is explicit and detailed (dedup, tiering, synthesis, curation cull), but there is no 'Use when...' clause or equivalent trigger guidance, which caps completeness at 3 per the judging guidelines. Not 2: the 'what' half is strong, not vague.	3 / 5
Trigger Term Quality	Relevant keywords ('concept', 'synthesize', 'cull', 'intellectual map') are present, but the natural phrases a user would actually say ('clean up my notes', 'dedupe my concepts', 'which ideas keep coming back') live only in the frontmatter triggers, not the description, and common synonyms are missing. Not 4: keyword coverage in the description itself is thin.	3 / 5
Distinctiveness Conflict Risk	Clearly niched to concept curation within a personal-brain corpus, with distinctive tier vocabulary ('T1 Canon to T4 Riff'). Not 5: it shares the concept-page domain with the ingestion skills it names (signal-detector, idea-ingest), leaving minor overlap risk with sibling stub-creation skills.	4 / 5
	Total	14 / 20 Passed

Validation

81%

Warnings & errors only

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

Validation — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (514 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
relative_links	Relative link issues: 5 missing, 5 suspicious	Warning

	Total	13 / 16 Passed

Repository: garrytan/gbrain
Path: skills/concept-synthesis/SKILL.md
Commit: e78f1c3

Reviewed: 5 days ago

Table of Contents

Discovery Implementation Validation

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.