hermes-agent

Expert in building self-improving AI agents with tool use, multi-platform messaging, and a closed learning loop. Proficient in LLM orchestration, tool integration, session management, and agent autonomy.

Quality

52%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Critical

Do not install without reviewing

Fix and improve this skill with Tessl

tessl review fix ./.agents/skills/hermes-agent/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 body is a comprehensive, well-organized project reference, but as a SKILL.md it is a monolithic 1700-line documentation dump with no progressive disclosure to separate files, padded with dated changelog and config-reference material, and lacking actionable task workflows with validation.

Suggestions

Move the bulk (release history, configuration reference, per-module architecture deep-dives) into separate reference files under references/ and keep SKILL.md as a lean overview with one-level-deep links, as the progressive-disclosure anchor expects.

Delete or relocate the dated release-history changelog — time-sensitive version/dating content is a named anti-pattern and inflates the token budget without aiding task execution.

If the skill is meant to guide building/extending the agent, add a concrete numbered workflow with validation checkpoints (e.g. build -> run tests -> verify) instead of descriptive internal-flow narratives.

Dimension	Reasoning	Score
Conciseness	At ~1765 lines the body is a full project manual, and sections like the dated release-history changelog (~85 lines), full config-reference tables, and end-to-end architecture deep-dives are padded reference material that does not belong in a lean SKILL.md; it is not score-1 because it does not explain basic concepts Claude already knows.	2 / 5
Actionability	There is genuine concrete content (copy-paste install commands, `hermes` CLI commands, real Python signatures and the agent loop), but large portions are descriptive prose and tables documenting an existing codebase rather than executable task guidance, so it sits at the middle anchor.	3 / 5
Workflow Clarity	Numbered sequences exist (e.g. the compression algorithm, message flow, context-file discovery) and are reasonably clear, but they describe internal behavior rather than an instructional workflow and lack any validation/feedback checkpoints, which the rubric caps at 3 for operations without validation.	3 / 5
Progressive Disclosure	No bundle directories (references/, scripts/, assets/) exist and there are zero external file references — content that clearly belongs in separate files (release history, config reference, full architecture) is entirely inlined into one monolithic document, despite a useful internal table of contents.	2 / 5
	Total	10 / 20 Passed

Description

66%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 specific and uses appropriate third-person voice with a recognizable niche, but it omits any explicit "Use when..." trigger guidance, which caps completeness and limits its value as a skill trigger.

Suggestions

Add an explicit 'Use when ...' clause naming concrete trigger phrases (e.g. 'Use when building or extending self-improving AI agents, wiring up tool use, or adding multi-platform messaging to an agent').

Tighten the trailing proficiency list ('Proficient in LLM orchestration, tool integration, session management, and agent autonomy') into concrete actions to lift specificity from 4 toward 5.

Include a couple of natural synonyms or file/keyword triggers users would actually say to push trigger_term_quality toward 5.

Dimension	Reasoning	Score
Specificity	Lists several specific capability areas — "tool use, multi-platform messaging, and a closed learning loop" plus "LLM orchestration, tool integration, session management, and agent autonomy" — which goes beyond 1-2 actions, though they read as proficiency domains rather than crisp performative actions like the 5-anchor examples.	4 / 5
Completeness	The "what" is clearly stated (building self-improving AI agents with the listed proficiencies), but there is no "Use when..." clause or equivalent trigger guidance, so per the rubric completeness is capped at 3 rather than reaching 4.	3 / 5
Trigger Term Quality	Natural terms such as "self-improving AI agents", "tool use", and "multi-platform messaging" would plausibly be said by a user, giving good coverage; it falls short of 5 because synonyms and file-extension-style triggers are absent.	4 / 5
Distinctiveness Conflict Risk	The "self-improving AI agents" + "closed learning loop" + "multi-platform messaging" framing carves a recognizable niche that is mostly distinct, though the broad enumeration of proficiencies leaves minor overlap risk with general agent-building skills, keeping it below 5.	4 / 5
	Total	15 / 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 (1766 lines); consider splitting into references/ and linking	Warning

	Total	15 / 16 Passed

Repository: fathah/hermes-desktop
Path: .agents/skills/hermes-agent/SKILL.md
Commit: 3aaadb0

Reviewed: 15 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.