flow-discover

Multi-AI research using available external providers (Double Diamond Discover phase)

Quality

38%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/flow-discover/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 execution contract core is genuinely actionable with strong validation gates, but it is buried in roughly 600 lines of duplication, conflicting legacy instructions, and inlined reference material. The two competing workflows and dangling file references make the skill hard to follow reliably despite its detailed gating.

Suggestions

Delete the duplicated legacy content: the second 'MANDATORY Context Detection & Visual Indicators' section, the 'How It Works' steps that call orchestrate.sh discover directly (they contradict the probe-single contract), and the repeated banner/Visual Indicators blocks.

Move banner templates, example presentations, Dev/Knowledge presentation formats, and the security-framing detail into references/ files linked one level deep from SKILL.md, and fix or remove the dangling references to skill-security-framing.md and skills/blocks/codex-host-adapter.md.

Define ${INTENSITY}, ${PROMPT}, ${codex_status}, and ${CLAUDE_SESSION_ID} at first use (or replace with literal instructions) so the bash snippets are truly copy-paste ready.

Dimension	Reasoning	Score
Conciseness	The body is ~855 lines with heavy padding: Steps 1-2 (context detection and banners) are duplicated nearly verbatim, banner templates appear four times, plus decorative ASCII art, a 'Benefits of hybrid approach' rationale section, and full example presentation dumps. This matches anchor 1 (severely verbose, heavily padded) rather than 2 because whole sections are redundant, not merely loose.	1 / 5
Actionability	Guidance is mostly executable: exact script invocations, run-id/nonce generation, fleet parsing rules, verification gates, and copy-ready bash blocks. Minor gaps keep it below 5: ${INTENSITY} and ${codex_status} are used before being defined, and Agent(...)/background_task(...) tool calls are host-pseudocode.	4 / 5
Workflow Clarity	The mandatory 7-step contract is well sequenced with explicit validation gates (research-verify, synthesis-file checks, an error-handling table), but a second, conflicting 'How It Works' workflow instructs invoking orchestrate.sh discover directly — exactly what the contract prohibits — making the effective sequence incoherent. This fits anchor 3 (sequence present but with significant gaps/incoherence) rather than 4.	3 / 5
Progressive Disclosure	A monolithic single file: banner templates, example presentations, presentation formats, and security-framing documentation are all inlined where they clearly belong in separate reference files, and the referenced paths 'skills/blocks/codex-host-adapter.md' and 'skill-security-framing.md' do not exist in any bundle directory. Matches anchor 2 (minimal structure, separate-worthy content inlined, references buried/nonexistent) rather than 3 because there is no actual bundle structure at all.	2 / 5
	Total	10 / 20 Passed

Description

37%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 identifies a genuine niche but is thin: it states one generic action and provides no usage triggers. A user asking for research would not reliably select this skill over general research skills based on this text alone.

Suggestions

Add a 'Use when...' clause naming concrete triggers, e.g., 'Use when the user asks to research or compare a topic across multiple AI providers, or wants multi-perspective/deep-dive research.'

List 2-3 concrete actions instead of the generic 'research', e.g., 'Dispatches parallel research probes to installed external AI providers (Codex, Copilot, Perplexity, etc.) and synthesizes their findings into a cited report.'

Include natural synonyms users would actually say (multi-AI, multi-provider, cross-model, research sweep) and drop or gloss the internal jargon 'Double Diamond Discover phase'.

Dimension	Reasoning	Score
Specificity	"Multi-AI research" names the domain but "research" is a single generic action; no concrete actions (e.g., which providers are dispatched, what output is produced) are stated. It fits anchor 2 (domain named, actions minimal/generic) rather than 3 because not even 1-2 concrete actions are enumerated.	2 / 5
Completeness	A recognizable "what" is present (multi-AI research via external providers) but there is no "when" clause at all — no "Use when..." guidance, so completeness is capped at 3 per the rubric guideline. Not 2 because the "what" is not extremely vague.	3 / 5
Trigger Term Quality	The only natural keyword is "research"; "Double Diamond Discover phase" and "external providers" are process jargon users would not naturally say, and common variations like "look into", "compare options", or "deep dive" are absent. Not 1 because "research" is a genuinely natural trigger term.	2 / 5
Distinctiveness Conflict Risk	"Multi-AI research using available external providers" carves out a niche but could still trigger for generic research, analysis, or comparison requests that other research skills would also serve. Not 4 because no distinct trigger phrases separate it from those neighbors.	3 / 5
	Total	10 / 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 (859 lines); consider splitting into references/ and linking	Warning

	Total	15 / 16 Passed

Repository: nyldn/claude-octopus
Path: skills/flow-discover/SKILL.md
Commit: 0093dab

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.