Tessl • The package manager for agent skills

This is an exemplary skill body: a lean, dense procedural workflow with concrete commands, explicit validation and falsification checkpoints, and a clean progressive-disclosure structure whose references all exist and deliver exactly what the body promises. No weaknesses found across any dimension.

Dimension	Reasoning	Score
Conciseness	The body is ruthlessly lean with zero concept explanations or padding — lines like "Cluster only on >=3 evidences. One anecdote is not a finding." and "Dead hypotheses die here, silently." are pure instruction. Every token earns its place; the skill assumes Claude's competence throughout.	5 / 5
Actionability	Every step carries concrete, executable guidance: parameterized commands ("ywai eval run --days <n> --worktree <w> --json", "GET /api/evals/runs"), with full copy-paste commands in references/queries.md, exact template slot IDs in references/report.md, and a decision table in references/thresholds.md — all verified to exist and deliver. Common cases across sessions, evals, memory, git, and Windows quirks are covered.	5 / 5
Workflow Clarity	Seven numbered steps (Scope → Collect → Validate → Cluster → Interrogate → Propose → Report) are clearly sequenced with explicit validation checkpoints (freshness check first, threshold gates in step 3) and a falsification feedback loop in step 5 ("write the disproof before the proposal... Dead hypotheses die here"). The Rules section doubles as a checklist, including the "One failed source never stops the run" error-recovery rule.	5 / 5
Progressive Disclosure	The body is a concise overview with well-signaled, one-level-deep references (queries.md, thresholds.md, analysis.md, report.md) plus assets/report-template.html — all verified to exist, none nested further. Each reference is linked at the exact step where it is needed, and detail is appropriately split out (command list, thresholds table, analysis method, report contract).	5 / 5
	Total	20 / 20 Passed

Description

61%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 distinct and has an explicit trigger clause, but it is overly terse: it names the domain without stating what the skill actually does (collect, cluster, falsify, propose, report) or what it produces. Trigger coverage is thin, leaning on one English keyword and one Spanish phrase. Adding concrete actions and natural trigger variations would lift specificity and completeness.

Suggestions

State 2-3 concrete actions and the output, e.g. "Collects session, eval, and memory data, clusters repeated frictions, and proposes one experiment per friction with an HTML report. Use when asked for a retro, retrospective, or weekly review of agent workflow pain points."

Add natural trigger variations users would actually say — "retrospective", "weekly review", "what went wrong this week" — alongside the existing "retro" and "auditar semana".

Mention the deliverable (HTML findings report) in the description so users know what to expect from invocation.

Dimension	Reasoning	Score
Specificity	"Weekly workflow retro over sessions, evals, memory" names the domain and data sources but contains no concrete action verbs (no "analyzes", "identifies", "generates") and never states the output. This matches the anchor 'Names the domain but actions are minimal or generic'; it falls below anchor 3 because not even 1-2 specific actions are listed.	2 / 5
Completeness	Both what ("Weekly workflow retro over sessions, evals, memory") and when ("Trigger: retro, auditar semana" — an explicit trigger clause equivalent to 'Use when') are present, so it clears the score-3 cap for missing trigger guidance. Not score 5: the 'what' omits actions and outputs, and the trigger list is thin and only partially in English.	4 / 5
Trigger Term Quality	"Trigger: retro, auditar semana" includes the natural keyword "retro", but misses common variations and synonyms ("retrospective", "weekly review", "postmortem") and relies on a niche Spanish phrase. Not score 4, since more than a few natural terms a user would actually say are absent.	3 / 5
Distinctiveness Conflict Risk	A clear niche — a weekly retrospective over internal session/eval/memory tracking data — with distinctive trigger terms, so the risk of firing for the wrong skill is minimal. The scoping to sessions/evals/memory data separates it from generic review or retrospective skills.	5 / 5
	Total	14 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: YoizenSA/dev-ai-workflow
Path: ywai/skills/workflow-retro/SKILL.md
Revision: 36f5af5

Reviewed: 5 days ago

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.

workflow-retro

Table of Contents

Is this your skill?