CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

flow-deliver

Multi-AI validation, scoring, and review using available external providers (Double Diamond Deliver phase)

Quality

60%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/flow-deliver/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

52%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 is highly actionable — concrete bash, explicit sequencing, validation gates, and error handling — but it is bloated and structurally confused: the same opening steps appear twice with conflicting banners, template placeholders are unresolved, and everything (including a 130-line example report) lives inline in one enormous file with no reference bundle. Consolidating the duplicated sections and moving examples/templates to reference files would materially improve it.

Suggestions

Merge the 'Deliver Workflow - Deliver Phase' section into the single EXECUTION CONTRACT so context detection and the banner appear exactly once, with one authoritative banner template.

Move the full Example 1 validation report and the 'Validation Report Format' template into a reference file (e.g. references/report-format.md) and keep only a short pointer in SKILL.md.

Resolve or substitute the {{PREAMBLE}}, {{VISUAL_INDICATORS}}, and {{QUALITY_GATES}} placeholders (or explain that the loader expands them) so the skill reads as complete when loaded directly.

DimensionReasoningScore

Conciseness

The ~870-line body duplicates its two mandatory opening steps in two separate sections (the 'EXECUTION CONTRACT' and the 'Deliver Workflow - Deliver Phase' section) with two conflicting banner templates, includes a ~130-line fully worked example report, ASCII-art diagrams, and repeated 'MANDATORY - CANNOT SKIP' admonitions. This matches 'Noticeably verbose; several unnecessary explanations or padded sections'; not 1 because it does not explain concepts Claude already knows, and not 3 because the duplication and padding are extensive, not incidental.

2 / 5

Actionability

Nearly every step ships copy-paste-ready bash with fallback chains ("${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}/.claude-octopus/plugin}}"), a run-marker validation pattern, a PR-posting script, and per-step error handling. It is not 5 because unresolved template placeholders ({{PREAMBLE}}, {{VISUAL_INDICATORS}}, {{QUALITY_GATES}}), literal '<user's validation request>' substitutions, and the two conflicting instruction blocks leave minor gaps the model must reconcile.

4 / 5

Workflow Clarity

Steps 1-7 are explicitly sequenced with blocking gates ('DO NOT PROCEED TO STEP N until...'), a dedicated validation gate (Step 5 rejects missing/failed reports), per-step error handling, and a completion checklist. Not 5 because the duplicated second presentation of the workflow (its own Steps 1-5 with a different banner and no provider check) contradicts the EXECUTION CONTRACT and could derail execution; not 3 because validation checkpoints and error-recovery guidance are otherwise explicit and strong.

4 / 5

Progressive Disclosure

The skill is an 870-line monolith with no bundle reference files (references/, scripts/, assets/ are absent), and large blocks that clearly belong in separate files — the full example validation report, the report format template, the banner templates — are inlined, twice in the banners' case. This fits 'content that clearly belongs in separate files is inlined'; not 3 because beyond the missing split, the duplication makes navigation actively confusing, and no actual one-level-deep reference documents exist in the bundle.

2 / 5

Total

12

/

20

Passed

Description

50%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 is serviceable but mediocre: it says what the skill does at a generic level while omitting when to use it and what it operates on. Adding an explicit 'Use when...' clause with concrete trigger phrases (e.g., 'validate code before shipping', 'review a PR with multiple AI providers') would raise both completeness and trigger-term quality.

Suggestions

Add an explicit trigger clause, e.g. 'Use when asked to validate or review code/documents before shipping, or when multi-provider (Codex + Antigravity + Claude) review is wanted'.

Name the objects of the actions — code, documents, PRs — so 'validation, scoring, and review' become concrete capabilities.

Include natural synonyms users would say (code review, security audit, quality check) to improve trigger matching and distinctiveness.

DimensionReasoningScore

Specificity

The description names its domain ("Multi-AI validation, scoring, and review using available external providers") and lists a few actions — validation, scoring, review — but the actions are generic verbs with no objects (code, documents, PRs). This matches the anchor 'Names domain and 1-2 concrete actions, but not comprehensive'; it is not a 2 because three distinct actions are enumerated, and not a 4 because none are concrete enough to tell a user what artifacts get validated.

3 / 5

Completeness

The 'what' is stated (multi-AI validation, scoring, and review via external providers) but there is no 'Use when...' clause or equivalent trigger guidance anywhere in the description, capping completeness at 3 per the judging guidelines. It is not a 2 because the 'what' is clear rather than vague; it is not a 4 because 'when' is entirely absent rather than merely implicit.

3 / 5

Trigger Term Quality

Terms like "validation", "scoring", and "review" are words users would naturally say, but common variations ("code review", "audit", "quality check", "security review") and any file/extension hooks are absent, and "Double Diamond Deliver phase" is internal jargon. This fits 'Some relevant keywords but missing common variations or synonyms' rather than 4, where only a few natural terms would be missing.

3 / 5

Distinctiveness Conflict Risk

The phrase "Multi-AI validation... using available external providers (Double Diamond Deliver phase)" carves out a niche distinct from single-model review skills, but a plain request like "review this code" would plausibly match it and sibling review skills alike. This is 'Somewhat specific but could still overlap with similar skills' — not 4, since no distinct trigger phrases separate it from generic review/validation skills.

3 / 5

Total

12

/

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 (889 lines); consider splitting into references/ and linking

Warning

Total

15

/

16

Passed

Repository

nyldn/claude-octopus

Path

skills/flow-deliver/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.