Tessl • The management layer for AI agent skills

SkillPublic

autoresearch

Stateful single-mission improvement loop with strict evaluator contract, markdown decision logs, and max-runtime stop behavior

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 ./skills/autoresearch/SKILL.md

Quality

Content

67%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 compact, well-structured instruction-only skill with a clear sequence and explicit stop conditions, and it wisely avoids teaching known concepts. Its main weakness is actionability: the evaluator invocation, evaluation JSON shape, and decision-log entry format are all left unspecified, so Claude must improvise the actual execution details.

Suggestions

Add a concrete example evaluation JSON (e.g., {"pass": false, "score": 0.62, "summary": "..."}) so the required structured output is unambiguous.

Show a minimal decision-log entry template (timestamp, iteration, change made, evaluation outcome, next decision) to make step 3 of the workflow directly executable.

Specify how the evaluator reference is resolved and run (command line, script path, expected exit behavior) instead of the abstract 'run the evaluator'.

Dimension	Reasoning	Score
Conciseness	The body is lean, uses compact tagged sections, and assumes competence (no basic-concept explanations), but the Contract, Do_Not_Use_When, and Execution_Policy sections each restate the same v1 single-mission and 'do not use the deprecated CLI' points, so minor trimming is possible.	4 / 5
Actionability	The directory layout and artifact list are concrete, but there are no executable commands or examples for the key details: how the evaluator is invoked, what fields the evaluation JSON must contain, or what a decision-log entry looks like — leaving the guidance incomplete rather than copy-paste ready.	3 / 5
Workflow Clarity	The Workflow section gives a clear numbered sequence with a per-iteration evaluator checkpoint and explicit stop conditions, forming an implicit feedback loop; the gap is that no recovery path is defined when the evaluator itself errors (vs. merely returning non-pass).	4 / 5
Progressive Disclosure	No bundle files exist and the body is well organized into distinct tagged sections with an inline artifact-shape example, but at roughly 78 lines it exceeds the simple-skill threshold and inlines detail (artifact tree, cron integration) that could live in a reference file.	4 / 5
	Total	15 / 20 Passed

Description

46%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 conveys a distinctive, niche capability but is feature-oriented rather than action-oriented and entirely lacks 'when to use' trigger guidance. Natural user phrasing is largely absent in favor of internal jargon.

Suggestions

Add an explicit 'Use when...' clause naming natural user phrases (e.g., 'use when the user asks to iteratively improve a mission until an evaluator passes, or to keep iterating on experiments with logged decisions').

Replace jargon-heavy feature nouns with concrete action verbs (e.g., 'Iterates one mission through evaluator runs, records decisions as markdown logs, and stops at a max-runtime ceiling').

Include common user synonyms such as 'research loop', 'iterate until it passes', or 'experiment logging' to improve trigger matching.

Dimension	Reasoning	Score
Specificity	Names the domain ('single-mission improvement loop') and a couple of concrete features ('markdown decision logs', 'max-runtime stop behavior'), but describes characteristics rather than listing several specific actions, so it sits at the 3 anchor rather than 4.	3 / 5
Completeness	The 'what' is fairly clear (a stateful, evaluator-driven improvement loop with durable logs), but there is no 'use when' or equivalent trigger guidance at all, which caps completeness at 3.	3 / 5
Trigger Term Quality	Terms like 'strict evaluator contract', 'max-runtime stop behavior', and 'stateful single-mission' are technical jargon users would not naturally say; only 'improvement loop' is a semi-natural phrase, matching the one-or-two-generic-keywords anchor rather than the relevant-keyword-set of 3.	2 / 5
Distinctiveness Conflict Risk	The niche framing (single-mission, evaluator-driven, stateful, max-runtime stop) is mostly distinct with only minor overlap risk against generic research/iteration skills; not a 5 because it offers no distinct trigger phrases to disambiguate.	4 / 5
	Total	12 / 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
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: Yeachan-Heo/oh-my-claudecode
Path: skills/autoresearch/SKILL.md
Revision: f2dffcc

Reviewed: 3 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.

autoresearch

Table of Contents

Is this your skill?