grasp

Verify understanding of a target in play — code, a document, a result — present in context and quotable. Type: (TargetUngrasped, User, VERIFY, Target) → VerifiedUnderstanding

Quality

42%

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 ./katalepsis/skills/grasp/SKILL.md

Quality

Content

35%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 skill documents a complex comprehension-verification protocol with explicit phasing and checkpoints, but it is severely dense and abstract, restates its core rationale redundantly, and provides almost no concrete enactable examples. It would be far more actionable if lean and exemplified.

Suggestions

Replace or supplement the formal-specification block with a concrete worked example of a /grasp round (sample entry-point labels, a sample probe, a sample adjudicated answer with an attached excerpt).

Cut the redundant restatement of the adjudication/basis-attachment rationale to a single canonical statement in Rules, removing the repeats in Definition, Phase 3, and Tool Grounding.

Split the large formal type/morphism block into a references file (e.g. references/protocol-spec.md), keeping SKILL.md a lean operational overview with the one-level-deep reference.

Dimension	Reasoning	Score
Conciseness	The body is noticeably verbose: a single large formal-specification code block carries dense type algebra and long inline '--' annotations, and the adjudication/basis-attachment rationale is restated across Definition, Phase 3, Tool Grounding, Protocol, and Rules, fitting the 'noticeably verbose; several unnecessary explanations or padded sections' anchor; it avoids explaining concepts Claude already knows, which keeps it above 1.	2 / 5
Actionability	Guidance is largely abstract ('Derive up to three first-turn labels...', 'Present the selected artifact context and a concrete scenario before each probe') with almost no concrete enactable examples — no sample probe phrasing, no example dialogue, no concrete round — so a model gets high-level hints but is missing the specific steps to execute, fitting the 'minimal concrete guidance' anchor.	2 / 5
Workflow Clarity	Phases 0–3 and the loop are explicitly sequenced with explicit checkpoints (ZeroGapConfirmation, coverage routing, admissible(HC) guard, terminal shapes), but the extreme density and repeated formal restatement bury the checkpoints and make the operational sequence hard to follow, fitting the 'steps listed with sequence present but checkpoints obscured' anchor.	3 / 5
Progressive Disclosure	There is one real, clearly signaled, one-level-deep reference (references/round-composition.md) with a concrete read-trigger, and section headers exist, but the bulk of the formal spec is inlined into a single monolithic code block rather than split out, fitting the 'some structure but content that should be separate is inline' anchor.	3 / 5
	Total	10 / 20 Passed

Description

50%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 clearly states what the skill does (verify understanding of a quotable target in context) but omits explicit 'Use when' trigger guidance and natural user-facing trigger phrases, capping completeness and trigger quality. It is adequately specific and reasonably distinct but would benefit from explicit activation triggers.

Suggestions

Add an explicit 'Use when...' clause naming concrete user triggers (e.g. 'Use when the user wants to confirm or firm up their understanding of code, a document, or a result already in context').

Surface natural trigger terms a user would actually say ('I don't understand this code', 'help me verify my understanding', '/grasp') rather than only formal vocabulary.

Name a second concrete action beyond 'verify' (e.g. probe, surface gaps, attach grounding excerpts) to lift specificity above 3.

Dimension	Reasoning	Score
Specificity	Names the domain (verifying understanding of a code/document/result target) and one concrete action ('Verify understanding') plus a precondition ('present in context and quotable'), but does not enumerate multiple specific concrete actions, fitting the 'names domain and 1-2 concrete actions' anchor.	3 / 5
Completeness	The 'what' (verify understanding of a target in play) is clear, but there is no explicit 'Use when...' clause or equivalent trigger guidance, which per the rubric caps completeness at 3; 'when' is only weakly implied via the quotable-in-context precondition.	3 / 5
Trigger Term Quality	Contains relevant terms (understand, code, document, result) but lacks the natural phrases a user would actually say (e.g. 'I don't understand this', 'grasp', 'verify my understanding') and omits the /grasp invocation trigger, fitting the 'some relevant keywords but missing common variations' anchor.	3 / 5
Distinctiveness Conflict Risk	The comprehension-verification niche and Stoic/type-signature vocabulary are distinctive, yet 'understanding of code, a document, a result' is broad enough to overlap with explanation, review, and analysis skills, fitting the 'somewhat specific but could still overlap' anchor.	3 / 5
	Total	12 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: jongwony/epistemic-protocols
Path: katalepsis/skills/grasp/SKILL.md
Commit: 3ffe94c

Reviewed: 1 day 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.