contextualize

Detect application-context mismatch after execution. Fires when correct output may not fit the actual context. Type: (ApplicationDecontextualized, AI, CONTEXTUALIZE, Result) → ContextualizedExecution

Quality

43%

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 ./epharmoge/skills/contextualize/SKILL.md

Quality

Content

45%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 specifies a richly sequenced protocol with explicit validation gates and feedback loops, but it is a severely dense, monolithic formal specification that restates its invariants repeatedly and inlines material that should live in separate reference files. It is actionable in its Protocol section yet abstract and hard to navigate overall.

Suggestions

Move the large formal blocks (TYPES, FLOW, CONVERGENCE, TOOL GROUNDING, MODE STATE, COMPOSITION) into separate reference files and keep SKILL.md as a concise overview that links to them one level deep.

Cut repeated restatements of the same invariants (e.g., certificate-before-registration, Kind -> resolution form -> repair locus) so each is stated once where it is load-bearing.

Lead with a short, plain-language step-by-step checklist of the protocol before the formal detail, so a reader can follow the sequence without parsing the dense formal blocks.

Dimension	Reasoning	Score
Conciseness	The 580-line body is noticeably verbose with pervasive restatement of the same invariants (e.g., 'Kind -> resolution form -> repair locus', certificate-before-registration) across FLOW, TYPES, PHASE TRANSITIONS, LOOP, CONVERGENCE, and COMPOSITION, adding many padded explanatory asides; not a 1 because the formalism is novel rather than concepts Claude already knows.	2 / 5
Actionability	The Protocol section provides concrete, near-copy-paste gate renderings and rendering instructions, but the bulk of the body is abstract formalism and pseudocode-style FLOW/type definitions rather than executable guidance, so it sits at some-concrete-but-incomplete rather than mostly-executable.	3 / 5
Workflow Clarity	The phased sequence (Phase 0 -> Phase 1 -> Phase 2 -> LOOP -> CONVERGENCE) is explicit with real validation checkpoints (fail-closed certificate gate, Qz zero-mismatch confirmation, Qa attribution, convergence trace) and an Adapt re-scan feedback loop; it is not a 5 because the extreme density and branching complexity prevent a clean, digestible checklist.	4 / 5
Progressive Disclosure	Despite section headers and one real, clearly-signaled reference (references/round-composition.md), the vast majority of the formal specification (TYPES, FLOW, CONVERGENCE, TOOL GROUNDING, MODE STATE) that clearly belongs in separate reference files is inlined into a monolithic 580-line SKILL.md, fitting the minimal-structure/inlined-content anchor.	2 / 5
	Total	11 / 20 Passed

Description

42%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 conveys a clear niche concept and a firing condition, but it is written in abstract formal jargon with a single concrete action and no natural 'Use when...' trigger guidance. It is distinguishable but thin on actionable specifics and user-natural keywords.

Suggestions

Add a concrete 'Use when...' clause with natural trigger phrases a user would actually say (e.g., 'Use when a result is technically correct but may not fit the user's actual situation, constraints, or environment').

List the concrete actions the skill performs beyond detection (surface mismatches one at a time, gather user judgment, adapt or withdraw the result), not just 'Detect'.

Replace or supplement the typed signature jargon with plain-language keywords and synonyms to improve trigger-term quality.

Dimension	Reasoning	Score
Specificity	The description names the domain ('application-context mismatch after execution') and one concrete action ('Detect'), but offers no further concrete actions such as surfacing, judging, or adapting the result, so it stops at the domain-plus-1-2-actions anchor rather than the multi-action anchor.	3 / 5
Completeness	It states a clear 'what' (detect application-context mismatch after execution) and a weak 'when' ('Fires when correct output may not fit the actual context'), but there is no 'Use when...' clause with concrete trigger phrases, which the rubric caps at 3.	3 / 5
Trigger Term Quality	The keywords ('application-context mismatch', 'correct output may not fit the actual context') are abstract jargon rather than natural phrases a user would utter, and there are no synonyms or concrete trigger terms; this fits the one-or-two-generic-keywords anchor better than the some-relevant-keywords anchor.	2 / 5
Distinctiveness Conflict Risk	The niche framing (post-execution applicability verification, the typed signature) is somewhat specific, but the abstract language could still overlap with general review or critique skills, fitting the somewhat-specific-but-overlap anchor rather than the mostly-distinct one.	3 / 5
	Total	11 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (580 lines); consider splitting into references/ and linking	Warning

	Total	15 / 16 Passed

Repository: jongwony/epistemic-protocols
Path: epharmoge/skills/contextualize/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.