flashrag-evidence

Local evidence retrieval (FlashRAG-style) for VCO/vibe: search protocols/config/skills docs and return citeable snippets with file+line anchors.

Quality

64%

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 ./bundled/skills/flashrag-evidence/SKILL.md

Quality

Content

75%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 a well-organized, actionable overview with concrete commands, a coverage-driven feedback loop, and a real one-level-deep script reference. The main weaknesses are hardcoded user-specific paths limiting copy-paste reuse and slightly explanatory asides that could be tightened.

Suggestions

Parameterize or templatize the hardcoded user paths (C:\Users\羽裳\...) so commands are copy-pasteable across machines, e.g. use a $SKILL_DIR or ~/.codex/skills/flashrag-evidence variable.

Tighten the explanatory asides (the removed install-upstreams.ps1 rationale and the preflight non-secrets note) to one clause each to push conciseness toward fully lean.

Add an explicit one-line 'See scripts/flashrag_evidence.py for CLI flags' navigation pointer to make the script reference clearly signaled rather than only embedded in command examples.

Dimension	Reasoning	Score
Conciseness	Lean, well-sectioned body with no concept-explanation padding (no lecture on what BM25 is), though a few explanatory asides like the removed install-upstreams note could be trimmed, keeping it just below the fully-lean 5.	4 / 5
Actionability	Provides concrete, mostly copy-paste-ready commands with real args (--query, --topk, --engine bm25s) and an rg fallback, but hardcoded user-specific Windows paths (C:\Users\羽裳\...) prevent fully generic copy-paste execution.	4 / 5
Workflow Clarity	Clear 4-step sequence with a feedback loop in step 4 (low coverage -> expand roots/topk -> rg fallback), but coverage-based checking is not an enforced correctness checkpoint and the preflight is an opt-in manual step, so it stops short of explicit validation-driven flow at 5.	4 / 5
Progressive Disclosure	Clean section structure with a real one-level-deep script reference (scripts/flashrag_evidence.py) that is signaled by command path and verified to exist, but references are referenced as command paths rather than explicit 'See X for details' navigation, leaving minor organization gaps relative to the 5 anchor.	4 / 5
	Total	16 / 20 Passed

Description

53%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 clearly states what the skill does and carves a specific niche, but it lacks an explicit 'Use when...' trigger clause and relies on domain-specific jargon over natural user phrasing. It is a solid, slightly above-midpoint description with room to add explicit trigger guidance.

Suggestions

Add an explicit 'Use when ...' clause naming concrete user phrases (e.g., 'Use when you need to cite where a VCO rule or threshold comes from in local docs/config').

Include natural-language synonyms alongside the technical terms (e.g., 'find the source of a rule', 'quote the exact doc snippet') so trigger matching catches general phrasing.

Optionally add a one-line scoping note in the description itself that this is local-docs-only (distinct from web/code search), since that boundary currently lives only in the body.

Dimension	Reasoning	Score
Specificity	Names the domain and two concrete actions ('search protocols/config/skills docs' and 'return citeable snippets with file+line anchors'), but it does not list several actions so it falls short of the 'several specific actions' bar for a 4.	3 / 5
Completeness	Provides a clear 'what' but no explicit 'Use when...' trigger clause; the 'when' is only weakly implied by 'for VCO/vibe', which per the rubric guideline caps completeness at 3.	3 / 5
Trigger Term Quality	Contains relevant domain keywords (evidence retrieval, citeable snippets, file+line anchors) but leans on technical/VCO jargon and omits the natural phrases a general user would say, missing common synonyms.	3 / 5
Distinctiveness Conflict Risk	Carves a distinct niche (local docs/config evidence retrieval differentiated from code-graph and web tools) with minor overlap risk, fitting 'mostly distinct' rather than the fully differentiated 5.	4 / 5
	Total	13 / 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
referenced_paths_exist	Referenced path issues: 1 missing	Warning

	Total	15 / 16 Passed

Repository: foryourhealth111-pixel/Vibe-Skills
Path: bundled/skills/flashrag-evidence/SKILL.md
Commit: d5ae560

Reviewed: 1 day 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.