Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add jiangyuxue666/skill-auto-maintain --skill skill-auto-maintaingit clone --depth 1 https://github.com/jiangyuxue666/skill-auto-maintainWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/jiangyuxue666/skill-auto-maintain/skill-auto-maintain)<a href="https://agentmods.dev/skills/jiangyuxue666/skill-auto-maintain/skill-auto-maintain"><img src="https://agentmods.dev/badge/skills/jiangyuxue666/skill-auto-maintain/skill-auto-maintain/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/jiangyuxue666/skill-auto-maintain/skill-auto-maintain"><img src="https://agentmods.dev/badge/skills/jiangyuxue666/skill-auto-maintain/skill-auto-maintain.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00069 | $0.01509 |
| Opus 5.5 | $0.00028 | $0.00604 |
| Sonnet 5.5 | $0.00014 | $0.00302 |
| Haiku 4.5 | $0.00007 | $0.00151 |
Grade A, and why
skill-auto-maintain scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 145 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill Auto Maintain
A standalone tool for managing non-bundled Hermes skills — works on any standard Hermes Agent installation without additional framework dependencies.
What It Does
Scans all skill directories under ~/.hermes/skills/ and runs four phases:
| Phase | Description |
|---|---|
| [Scan] | Identifies non-bundled skills, migrates misplaced ones to user_skills/, registers them in user_skills.json, marks registry entries as deleted when their skill directory no longer exists on disk |
| [Compare] | Pairwise multi-dimensional similarity scoring to detect merge candidates |
| [Optimise] | Fixes SKILL.md formatting and checks content quality (triggers, tags, description, structure) |
| [Report] | Generates a comprehensive professional summary of all actions taken |
How to Execute
This skill requires running maintain.py via the Hermes venv Python. When the user asks to run skill auto maintenance:
- Determine the script path:
- First run (before self-migration):
~/.hermes/skills/skill-auto-maintain/maintain.py - After self-migration (subsequent runs):
~/.hermes/skills/user_skills/skill-auto-maintain/maintain.py - Try the
user_skills/path first since most users will already have run it once
- First run (before self-migration):
- Run the script:
~/.hermes/hermes-agent/venv/bin/python <script_path> - Return the full report output to the user — it contains all scan, migration, comparison, and optimisation results.
Note on self-migration: The first run automatically moves the tool's own directory from
skills/skill-auto-maintain/touser_skills/skill-auto-maintain/. This is expected — after one run, the tool lives inuser_skills/permanently. Subsequent runs should use theuser_skills/path.
Scan Scope
The tool scans all directories under ~/.hermes/skills/ — including user_skills/ itself. This means:
- New skills placed directly into
user_skills/get auto-detected and registered on the next run. - Skills deleted from
user_skills/get their registry entry marked asdeleted(not removed — preserves audit trail). - The tool's own directory (
skill-auto-maintain/) is a regular skill directory and gets processed like any other.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 145 lines · 69 tokens per session scan A 10eb20567fc3
skill-auto-maintain is a skill published in the GitHub repository jiangyuxue666/skill-auto-maintain (5 stars, last pushed 5mo ago), licensed MIT. It adds 69 tokens to every session and 1,509 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
hermes-evolution-orchestrator
Run one bounded ESRA review after a major architecture or skill change, repeated failure, or an explicit full-cycle request. Select only the meta-skills justified by observable evidence; skip routine successful work. Triggered by "orchestrate evolution", "run full ESRA cycle", "improve self-development cycle", "hermes…
esra-runtime
Activate esra-runtime when locating, explaining, or running ESRA Python tools (evolutionhook, skillvalidator, evolutiondashboard, experimentrunner, esraexport, hermesintegration, baselinemetrics, esralogger, humanoversight). Documents the installed package path under Hermes home. Triggered by ESRA tools, where are…
github-actions-integrator
Activate github-actions-integrator when the task involves GitHub Actions workflows, CI/CD pipelines, triggering builds, checking workflow status, analyzing build results, or integrating GitHub Actions into the self-development process. Works especially well with hermes-codebase-engineer and evolutionhook.py. Triggered…
loop-auditor
Activate loop-auditor to periodically review and improve the Self-Evolution Loop itself. Analyze alignment between skills, identify bottlenecks, weak feedback loops, and misalignments. Propose structural improvements to the development system. Works with self-improver and mental-model-updater. Triggered by loop…
ooda-framework
Activate ooda-framework when structuring decision-making, rapid adaptation under uncertainty, or enhancing any loop (learning, improvement, agent cycles). Use to make Observe-Orient-Decide-Act explicit in self-evolution, agent tasks, or meta-processes. Combines with loop-auditor, self-observer, experimenter…
hermes-codebase-engineer
Activate hermes-codebase-engineer for any programming, code analysis, integration, refactoring or development tasks related to Hermes Agent fork, skill integration, OODA implementation, or building ESRA features. Use when the task involves reading/writing code, creating integration scripts, adapting skills to Hermes…