contextualize

Detect application-context mismatch after execution. Verifies applicability when correct output may not fit the actual context, registering each mismatch through a fail-closed deficit-fit certificate before adaptation, producing contextualized execution. The transformative revalidation loop is non-monotone — adapting the result mutates the evaluation target and can breed emergent mismatches; re-scan is mandatory. This is the non-monotone/transformative side of the contextualize ↔ distill boundary (distill is the claim-preserving side: it may add resolution or excise released bookkeeping in place, but never mutates what the record asserts). Type: (ApplicationDecontextualized, AI, CONTEXTUALIZE, Result) → ContextualizedExecution. Alias: Epharmoge(ἐφαρμογή).

Quality

31%

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

27%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is a rigorous, internally consistent formal specification of a post-execution applicability-check protocol with concrete tool grounding and clear phases, but it is far too verbose and monolithic: a single enormous formal code block plus repetitive prose re-derives the same invariants many times, with no progressive disclosure into referenced files. It would score much higher if split into overview + reference files and tightened.

Suggestions

Move the full TYPES/MORPHISM/CONVERGENCE formal block into a separate references/ file and keep SKILL.md as a concise overview that links to it one level deep.

Deduplicate the repeated explanations of certificate-before-registration and the non-monotone transformative-revalidation loop — state each invariant once and reference it.

Convert the formal predicates and type signatures that a runtime executor needs into concrete, copy-paste-ready checklists/procedures so the operational contract is actionable without parsing the notation.

Dimension	Reasoning	Score
Conciseness	The body is an extreme wall of dense formal-specification text (a single ~165-line code block of type definitions plus long prose) repeatedly restating the same invariants across FLOW/MORPHISM/PHASE TRANSITIONS/LOOP/CONVERGENCE/Rules sections; it explains the same mechanics (certificate-before-registration, non-monotone revalidation, distill boundary) many times over, far exceeding what Claude needs and matching the anchor-1 'verbose; padded with unnecessary context'.	1 / 3
Actionability	It does provide concrete, executable operational guidance — real tool calls (TaskCreate/TaskUpdate/Edit/Write), explicit task-format templates, a surfacing-format template with exact phrasing, and a phased sequence — but much guidance is specified as abstract type-level invariants and predicates rather than copy-paste-ready instructions, and key behaviors are buried under formal notation, placing it between anchor-2 and anchor-3.	2 / 3
Workflow Clarity	Phases 0–2 are clearly sequenced with explicit checkpoints (Phase 0 silent checkpoint, Phase 1 surfacing, Phase 2 mutate-then-re-scan) and feedback loops (re-scan after adaptation, bounded re-assessment for ambiguous mismatches), but the surrounding formal-block density and repeated re-derivation of the same transitions obscure the actual sequence for a reader, and the heavy invariants add cognitive load the rubric penalizes.	2 / 3
Progressive Disclosure	There are no bundle files (references/scripts/assets are absent) and the entire specification is one monolithic SKILL.md with a single deeply nested formal code block; nothing is split into referenced files, matching the anchor-1 'monolithic wall of text; poor organization'.	1 / 3
	Total	6 / 12 Passed

Description

35%

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 states a clear purpose (detecting application-context mismatch after execution) but buries it under dense internal terminology and a formal type signature that no user would naturally invoke. It lacks an explicit 'Use when' trigger clause, capping completeness at 2, and its trigger terms are technical jargon rather than natural language.

Suggestions

Add an explicit 'Use when...' trigger clause naming natural user phrases (e.g. 'Use when a result is technically correct but may not fit the actual project context, conventions, environment, or audience').

Replace or front-load the internal vocabulary (ApplicationDecontextualized, fail-closed deficit-fit certificate, transformative revalidation) with plain-language actions a user would say.

Move the formal type signature and the distill-boundary contrast into the body; keep the description to the concrete capability and its triggers to improve distinctiveness.

Dimension	Reasoning	Score
Specificity	It names a concrete domain (application-context mismatch after execution) and several actions (verifies applicability, registers mismatches, adapts), but the actions are abstract process steps rather than the multiple discrete concrete operations the anchor-3 example lists, and they are wrapped in heavy jargon (deficit-fit certificate, transformative revalidation loop).	2 / 3
Completeness	It clearly answers 'what does this do', but 'when should Claude use it' is only implied (after execution, when correct output may not fit context) — there is no explicit 'Use when...' clause or equivalent trigger guidance, which the rubric caps at 2.	2 / 3
Trigger Term Quality	There are no natural user-facing keywords; the description is dominated by technical/internal vocabulary (ApplicationDecontextualized, fail-closed deficit-fit certificate, contextualize ↔ distill boundary, Epharmoge/ἐφαρμογή) rather than terms a user would naturally say, matching the anchor-1 'technical jargon' example.	1 / 3
Distinctiveness Conflict Risk	The niche (post-execution applicability/context mismatch vs a sibling 'distill') is fairly distinctive, but the description is so abstract and jargon-laden that it could plausibly overlap with adjacent skills and a user/model would struggle to tell it apart from generic 'review the result' skills, so it sits at the 'somewhat specific but could overlap' anchor.	2 / 3
	Total	7 / 12 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: epharmoge/skills/contextualize/SKILL.md
Commit: ab9aa56

Reviewed: 2 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.