CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

apportion

Use when the user is about to hand one goal to an autonomous run, or asks for it to be planned for one, and no plan yet says when each unit is done. The user takes a plan of units with checks, accepted gaps and held judgments, seeing what needed settling first and what returns at the end.

Quality

63%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./merismos/skills/apportion/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

48%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.

This is an intricately specified protocol whose workflow sequencing and failure handling are unusually thorough and whose single reference file is well managed. Its quality is dragged down by pervasive redundancy — the same rules restated in Definition, Protocol, CONVERGENCE, and Rules sections quadrupling the token cost — and by the total absence of worked examples, leaving the elaborate formal contract without a single concrete instantiation.

Suggestions

Eliminate the redundant restatements: the Rules section repeats the Protocol nearly verbatim and the handoff-obligations text appears three times — consolidate each to a single statement and reference it, which would roughly halve the body's token cost.

Add one worked example: a small stated goal cut into two units with concrete completion predicates, a rendered sheet excerpt, and the resulting emission entries, giving the model an instantiation of the abstract contract.

Move the ~600-line Lean formal block into references/ (loading it only when the formal contract is needed) and keep a short prose summary of its invariants in SKILL.md, so the always-loaded body is an overview rather than a monolith.

DimensionReasoningScore

Conciseness

The body restates the same material three to four times: the Definition paragraph mirrors the description, the sheet's contents appear in the `respond` docstring, the Protocol's 'The sheet' section, the CONVERGENCE block, and the Rules, while the `handoffObligations` text appears verbatim three times and the Rules section largely restate the Protocol. It avoids anchor 1's explaining of common knowledge — the Lean contract is the skill's own novel material — but the padding is pervasive and systematic, matching anchor 2's 'several unnecessary… padded sections' rather than the tighter anchor 3.

2 / 5

Actionability

The Protocol prose gives structurally specific instructions — sheet fields labelled by the question they answer, 'emit one entry per unit with its whole certificate', 'Park the packaged plan in one durable carrier and emit a navigation block' — and failure handling is concrete. But nothing is instantiated: no example of a completion predicate, a cut unit, or a rendered sheet, and the ~600-line Lean block is an elaborate specification rather than executable guidance, fitting anchor 3's 'some concrete guidance but incomplete; missing key details' rather than 4.

3 / 5

Workflow Clarity

The sequence is coherent and complete with explicit validation gates and error-recovery loops: unreadable handoff and too-thin scope relay out, a taking over a non-closable plan redraws the sheet saying what stands in the way, and an incomplete carrier write redraws with what was written and what is missing. It falls short of anchor 5 because the flow is dispersed across four restatements (FLOW comment, PHASE TRANSITIONS, the `apportion` definition, and Protocol prose) with no single checklist, making it harder to follow than a 5 requires.

4 / 5

Progressive Disclosure

The one bundle reference (references/round-composition.md) is real, one level deep, and well signalled with explicit load conditions ('Read references/round-composition.md before composing when…'), which is good practice. But the ~600-line Lean formal block is inlined monolithically in SKILL.md — content that clearly belongs in a separate reference file loaded on demand — matching anchor 3's 'content that should be separate is inline' rather than 4's 'most content is appropriately placed'.

3 / 5

Total

12

/

20

Passed

Description

62%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 has a strong, explicit, and distinctive 'when' clause targeting a narrow niche, and correctly uses third-person voice. Its weaknesses are on the 'what' side: the skill's actions are conveyed only through outcome nouns ('units with checks, accepted gaps and held judgments') rather than concrete verbs, and trigger coverage lacks synonyms or variations a user might naturally say.

Suggestions

State the skill's actions in third-person verb form up front (e.g. 'Apportions one stated autonomous goal into coarse execution units and derives each unit's completion conditions before the run begins') so the 'what' is explicit rather than inferred from the user's outcome.

Broaden trigger coverage with natural synonyms users would actually say, such as 'delegate this goal', 'run this while I'm away', or 'plan this for an unattended run', to catch phrasings beyond the single 'hand one goal to an autonomous run' formulation.

DimensionReasoningScore

Specificity

The description names the domain ('hand one goal to an autonomous run… no plan yet says when each unit is done') and deliverable elements ('a plan of units with checks, accepted gaps and held judgments'), but the skill's actions (cutting into units, deriving completion predicates, showing authority) are never stated as concrete verbs. It matches anchor 3 — domain plus 1–2 concrete actions — rather than 4, which requires a list of specific actions, and exceeds 2 because several concrete deliverable elements are named.

3 / 5

Completeness

Both halves are present: an explicit and specific 'Use when…' trigger clause, and a 'what' in the outcome sentence ('The user takes a plan of units with checks, accepted gaps and held judgments, seeing what needed settling first and what returns at the end'). The 'what' is outcome-framed and abstract rather than explicitly stating what the skill does, so it does not reach anchor 5's clear statement of both; it exceeds anchor 3 because the 'when' is explicit, not merely implied.

4 / 5

Trigger Term Quality

Phrases like 'hand one goal to an autonomous run', 'asks for it to be planned', and 'when each unit is done' are plausible user utterances for this domain, but no synonyms or common variations are covered (e.g. 'delegate', 'while I'm away', 'unattended', 'take over this goal'). This fits anchor 3 — some relevant keywords, missing common variations — rather than 4, whose fuller coverage of natural terms is absent.

3 / 5

Distinctiveness Conflict Risk

The trigger condition is a clear niche — handing one goal to an autonomous run with no existing condition-bearing unit plan — with minimal conflict risk against typical skills. The clause 'asks for it to be planned for one' carries minor overlap risk with generic planning/scheduling skills, keeping it at anchor 4 rather than 5.

4 / 5

Total

14

/

20

Passed

Validation

93%

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

CriteriaDescriptionResult

skill_md_line_count

SKILL.md is long (1050 lines); consider splitting into references/ and linking

Warning

Total

15

/

16

Passed

Repository

jongwony/epistemic-protocols

Path

merismos/skills/apportion/SKILL.md

Reviewed

Table of Contents

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.