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

42%

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

35%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The skill body is a well-structured but massively verbose single-file codebase tour rather than lean, actionable guidance, and it makes no use of progressive disclosure via separate bundle files. It earns a middle score on actionability/workflow/structure but fails conciseness outright.

Suggestions

Collapse the body to a lean overview and move exhaustive detail (release history, full config tables, per-module architecture, design-pattern catalog) into separate reference files under references/ that SKILL.md links to, restoring the context-window budget.

Replace descriptive codebase tours with concrete, copy-paste-ready instructions for the skill's actual purpose (building/improving an agent): e.g. a numbered workflow for registering a tool, wiring a messaging adapter, or adding a learning-loop step, each with executable commands.

Add explicit validation checkpoints to any multi-step or risky workflow (e.g. 'register tool → run pytest → verify schema appears in get_tool_definitions() before proceeding'), since the current body has no validate→fix→retry feedback loops.

Dimension	Reasoning	Score
Conciseness	The body is a 1765-line monolithic project guide that exhaustively documents an entire codebase (architecture, every module, release history, config tables) and explains concepts Claude already knows, which is the verbose, padded-with-unnecessary-context pattern of anchor 1.	1 / 3
Actionability	It contains concrete, executable material (install commands, the agent loop, registry API, SQL schema, constructor signatures), but the bulk describes the system rather than instructing how to build or improve an agent, so it provides some concrete guidance without being fully actionable task instruction.	2 / 3
Workflow Clarity	Content is organized into a sequenced Table of Contents (overview → architecture → ... → pitfalls) and a 'Known Pitfalls' section, but there is no multi-step build/improve workflow with explicit validation checkpoints or feedback loops, matching the anchor-2 'sequence present but checkpoints missing' case.	2 / 3
Progressive Disclosure	No bundle files exist (no references/, scripts/, or assets/ directories) and the entire content lives in one 1765-line SKILL.md; sections are well-organized but material that should be split out (release history, config reference, design patterns) is all inline with no one-level-deep references, fitting anchor 2.	2 / 3
	Total	7 / 12 Passed

Description

50%

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 clear domain and several relevant capabilities, but it is buzzword-heavy, lacks any explicit 'Use when' trigger, and is generic enough to overlap with other agent-building skills. All dimensions land at the middle anchor.

Suggestions

Add an explicit 'Use when...' clause stating the natural situations that should trigger this skill (e.g. 'Use when building or extending an AI agent, wiring up multi-platform messaging, or adding a self-improvement/learning loop').

Replace abstract phrases like 'LLM orchestration' and 'agent autonomy' with concrete actions a user would request, such as 'route LLM calls across providers', 'register custom tools', or 'persist and search agent sessions'.

Add natural trigger terms users actually say (e.g. 'AI agent', 'chatbot with tools', 'Telegram/Discord bot', 'agent memory') to improve distinctiveness and trigger coverage.

Dimension	Reasoning	Score
Specificity	Names the domain ('self-improving AI agents') and several concrete-ish areas ('tool use, multi-platform messaging, and a closed learning loop', 'LLM orchestration, tool integration, session management'), but these read more as capability categories than the specific concrete actions the anchor-3 example demands.	2 / 3
Completeness	It states what the skill covers but provides no 'Use when...' clause or equivalent trigger guidance, which the judging guidelines explicitly cap at 2; the 'when to use it' half is entirely missing.	2 / 3
Trigger Term Quality	It includes some terms a user might say ('AI agent', 'messaging', 'tool use'), but the phrasing is jargon-laden ('LLM orchestration', 'agent autonomy', 'closed learning loop') and misses the common natural variations, so it sits at anchor 2 rather than the full natural-keyword coverage of anchor 3.	2 / 3
Distinctiveness Conflict Risk	The description is somewhat specific to self-improving agents with messaging, but the broad 'AI agents / tool use / LLM orchestration' framing could overlap with many general agent-building skills, matching anchor 2 rather than a clear distinct niche.	2 / 3
	Total	8 / 12 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: 9bae718

Reviewed: 2 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.