CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

deepwork

High-cost orchestrator workflow for large, high-risk, multi-phase coding efforts with meaningful dependencies and review gates. Do not activate for routine multi-file changes.

Quality

82%

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

SKILL.md
Quality
Evals
Security

Quality

Content

92%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.

An exceptionally tight orchestration skill: every section issues concrete directives with real thresholds, exact file formats, and explicit validation and escalation loops, and it is correctly scoped to a single SKILL.md with no bundle needed. The only slack is light stylistic flourish in the painter analogy and a few aphoristic lines.

DimensionReasoningScore

Conciseness

The body is dense directive prose with no padding and assumes competence (never explains git, builds, or review concepts), but the 'Work Like a Painter' analogy (~15 lines) and passages like 'Work is never finished: there will always be more to harden' are stylistic flourishes that could be tightened. This is anchor 4 (efficient with minor trimming possible) rather than 5.

4 / 5

Actionability

For an instruction-only skill the guidance is maximally concrete: exact commands ('mkdir -p .slim/deepwork', '.gitignore' setup), precise file formats (head with 'status: active' and 'slug:', progress file 'at most about 40 lines', tombstone 'at most 15 lines'), hard thresholds ('about four or fewer' lanes, 'at most two re-reviews'), and a copy-paste-ready example ('Pass 2 gate — attempt 2 of 3'). Per the scoring note, absence of code is not penalized when guidance is this actionable.

5 / 5

Workflow Clarity

The full lifecycle is explicitly sequenced (Planning a Pass → Running a Pass → numbered Gate and Step Back → State → Completion) with validation checkpoints ('Run the broad validation once on the reconciled state', 'Accept a lane after checking one decisive artifact') and explicit error-recovery feedback loops ('After two misses from a lane, change the boundary', 'at most two re-reviews', re-plan at ~2x expected effort). This matches anchor 5.

5 / 5

Progressive Disclosure

There are no bundle files (references/, scripts/, assets/ are absent) and none are needed — this is self-contained orchestration guidance whose state-file conventions must stay inline to survive compaction. The body is organized into eight clear, well-navigable sections with no nested or buried references and nothing that clearly belongs in a separate file, matching the well-organized single-file ideal.

5 / 5

Total

19

/

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.

A distinctive, honest description with a clear niche and a useful negative boundary, but it describes the skill's character rather than its actions and lacks natural, affirmative trigger phrases. Adding concrete capabilities and a 'Use when...' clause with user-voiced terms would raise it substantially.

Suggestions

Add an affirmative trigger clause with natural user phrasing, e.g. 'Use when the user asks for a full subsystem build, an unsafe-to-partially-ship migration, a large refactor, or a multi-week coordinated effort' — this would lift completeness from 4 to 5.

Include 1-2 concrete actions users can expect, such as 'splits work into lanes, runs per-pass review gates, and tracks resumable state', to move specificity from 3 to 4.

Add common synonyms for the domain (migration, rewrite, greenfield build, architectural change) so trigger matching works on the words users actually say.

DimensionReasoningScore

Specificity

The description names the domain precisely ("large, high-risk, multi-phase coding efforts with meaningful dependencies and review gates") but states few concrete actions — it says what the skill is for rather than what it does (e.g., no verbs like 'plans lanes', 'runs gates', 'tracks state'). This matches anchor 3 (domain plus limited concrete actions) better than 4, which expects several specific actions listed.

3 / 5

Completeness

The 'what' is clear ("High-cost orchestrator workflow for large, high-risk, multi-phase coding efforts") and 'when' is addressed both positively via the target domain and negatively via "Do not activate for routine multi-file changes". It falls at anchor 4 ('when' present but could be more explicit) rather than 5 because there is no affirmative 'Use when...' clause with concrete trigger phrases, only an exclusion.

4 / 5

Trigger Term Quality

Phrases like "multi-phase coding efforts", "review gates", and "routine multi-file changes" are relevant but miss the natural terms a user would actually say for this need — 'migration', 'rewrite', 'greenfield build', 'refactor', 'big project'. Anchor 3 (some relevant keywords, missing common variations/synonyms) is the closest fit; it is above 2 because the keywords are domain-appropriate, not generic.

3 / 5

Distinctiveness Conflict Risk

The niche (high-cost, high-risk, multi-phase orchestration) is clearly carved out, and the explicit anti-trigger "Do not activate for routine multi-file changes" reduces conflict with ordinary coding skills. It is below 5 only because the positive triggers are abstract ('large, high-risk... efforts') and could still overlap with generic planning or orchestration skills.

4 / 5

Total

14

/

20

Passed

Validation

100%

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

alvinunreal/oh-my-opencode-slim

Path

src/skills/deepwork/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.