conduct

The work needs several lines of thinking, and their order, independence, combination, stopping point, or where results go is not obvious: settle the method before starting.

Quality

51%

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 ./hyphegesis/skills/conduct/SKILL.md

Quality

Content

38%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 contains a genuinely detailed interaction protocol with concrete presentation formats, but it is buried under hundreds of lines of abstract formalism and invented terminology that cost enormous context budget. A reader must reconstruct the workflow from scattered sections, and the lack of any worked example leaves the abstract machinery hard to apply.

Suggestions

Move the Lean formalism block into a reference file (e.g. references/conduct-contract.md) and keep a short summary of its key judgments in the body, cutting the main body to a fraction of its current size.

Add one worked example of a map: a small two-region move set with the slot table, one change ledger, and the resulting checkpoint list, so the abstract axes become concrete.

Consolidate the workflow into one linear step-by-step section (activation → map → gate → answer → ledger → convergence → handoff), leaving the Rules section only for constraints not already implied by the sequence.

Dimension	Reasoning	Score
Conciseness	The body is ~900 lines: a ~600-line Lean-style formal block ('namespace Hyphegesis' through 'end Hyphegesis') plus densely neologistic prose bury roughly 100 lines of actionable guidance in the Protocol section. Far more than minor trimming is needed, though it is not explaining well-known concepts verbatim.	2 / 5
Actionability	The 'User-facing realization' section gives concrete presentation formats (moves as an indented outline under regions, slot tables with 'old → new' changed cells, ledger lines as changed input → affected slot → consequence), but there is no worked example of a map and the Lean axioms are explicitly undetermined black boxes ('yours to make from the material in front of you'), so execution guidance is incomplete.	3 / 5
Workflow Clarity	A sequence exists (observe → map → gate → answer → ledger → sufficient → conduct trace → handoff) and Constitution gates/checkpoints are defined, but it is scattered across the FLOW comment, MORPHISM block, Mode Activation, Protocol prose, and Rules, requiring the reader to reconstruct the order.	3 / 5
Progressive Disclosure	Both references (references/decompose-recovery.md, references/round-composition.md) exist, are one level deep, and are well-signaled with read conditions. However, the giant inline Lean formalism block is content that clearly belongs in a separate reference file, leaving the main body a monolithic wall.	2 / 5
	Total	10 / 20 Passed

Description

41%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 functions well as a trigger condition — it states precisely when the skill applies — but it barely says what the skill does, and it uses abstract phrasing rather than terms users would naturally say. A reader can tell when to activate it but not what they will get.

Suggestions

Add one sentence of concrete 'what': e.g. 'Designs a conduct plan — move set, region cut, order, independence, reconciliation, termination, and routing — presented as one map the user corrects, then handed off for execution.'

Include natural trigger phrases users would say, such as 'plan the steps,' 'how should I approach this,' 'what order should I run these in,' or 'orchestrate these tasks.'

Adopt an explicit 'Use when...' construction so the trigger guidance reads as instruction rather than a descriptive condition.

Dimension	Reasoning	Score
Specificity	The description names the domain (multi-line reasoning whose order, independence, combination, stopping point, and routing are unclear) but the only action named is 'settle the method before starting' — minimal and generic. It lists no concrete capabilities, so it falls below the 1-2-concrete-actions anchor.	2 / 5
Completeness	The 'when' is explicit — the entire first clause states the trigger condition ('The work needs several lines of thinking, and their order... is not obvious') — but the 'what' is only the vague 'settle the method before starting,' which does not convey what the skill actually produces (a method plan, a map, checkpoints).	3 / 5
Trigger Term Quality	Phrases like 'several lines of thinking,' 'order, independence, combination, stopping point,' and 'settle the method' are abstract and technical; the natural words a user would actually say ('plan,' 'how should I approach,' 'workflow,' 'orchestrate,' 'steps') are missing.	2 / 5
Distinctiveness Conflict Risk	The niche is fairly distinct — deciding order/independence/reconciliation/termination/routing of multiple cognitive moves — with vocabulary unlikely to collide with file-format or tooling skills. Minor overlap risk remains with generic planning or task-management skills.	4 / 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 (992 lines); consider splitting into references/ and linking	Warning

	Total	15 / 16 Passed

Repository: jongwony/epistemic-protocols
Path: hyphegesis/skills/conduct/SKILL.md
Commit: 10abcb3

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.