Skip to content

rlaope/oh-my-hermes / skills/omh-docs

omh-docs

Verified

[omh] Explaining OMH itself: current-source-first documentation for OMH itself: product identity, public capability catalog, model routing, local state, and long-term memory. Use when the user says: product-docs, OMH documentation, oh-my-hermes documentation, what is OMH, what is oh-my-hermes, how does OMH work, OMH capability catalog, OMH skill catalog.

references/model-routing-and-local-state.md

View on GitHub
# Model Routing and Safe Local State Inspection

Load this reference for model-chain behavior, coding-owner routing, capability
policy, installed-profile questions, or where OMH stores local metadata.

## Public Routing Facts

For published behavior, read the current `MODEL_OPTI.md`, `docs/FANOUT.md`,
`docs/INSTALLATION.md`, relevant command help, and implementation at the
disclosed official ref. Treat documented defaults as versioned product facts,
not proof that this machine uses them.

## Passive Local Commands

Prefer these before opening files:

```sh
omh list --json
omh model-chains show
omh coding category-maestro show
omh capability-policy status
omh status
omh <command> --help
```

`omh doctor --json` is a diagnostic, not passive inspection: when its state
store is writable it records `last_doctor`, and failed health checks can return
a nonzero exit code while still emitting useful JSON. Disclose the
`state-write side effect` before running it, then interpret the payload rather
than treating a nonzero status as command absence. Use `--no-state-write` to
skip the state update and runtime/workflow write probes; those checks then
report write access as not probed and use a non-writing permission check. Use
`--no-personal-files` to avoid opening `SOUL.md`, `memories/MEMORY.md`, and
`memories/USER.md`; their two advisories are explicitly `skipped`. The flags
are independent and work with human or JSON output. The loader-registration
smoke can still use a temporary directory under `--no-state-write`.

Record the command, selected profile/home, and observed OMH version or commit.
Do not reinterpret a command's absence as a product-wide capability claim.

## Resolve the Active Home First

Use the current CLI metadata and `src/system/paths.py` to resolve exactly one
active home before opening a path. The user-scope default is `~/.omh`, but
`OMH_HOME` and `--omh-home` can override it, while `--scope project` resolves
to the repository's `.omh`. Do not probe several guessed homes.

Installer-managed skills can be served through an atomic generation pointer
such as `current/skills`; do not assume the active pack is a literal
`~/.omh/skills/` directory.

## Narrow Resolved-Home Fallback

Only when no passive command answers the question, inspect the minimum metadata
path needed under the one resolved home. Documented areas can include:

- `manifest.json` and `setup-profile.json`;
- the active managed skill generation and its `current/skills` pointer;
- `routing/` model chains, route provenance, provider entitlements,
  model-provider mappings, price overrides, dispatch defaults, and
  category-specific routing;
- target registries and goal, loop, state, or runtime metadata;
- memory and learning stores;
- generated operator artifacts described by the current docs.

Presence varies by OMH version, setup profile, enabled capability, scope, and
whether a workflow has ever produced that artifact. Report
`not_present_for_this_install` rather than treating every documented path as
mandatory.

## Secret and Privacy Boundary

Never read or print credentials, tokens, auth files, `.env` values, provider
secrets, raw private logs, raw prompts/transcripts, or unrelated user content.
Do not recursively list the home directory. Inspect filenames, schemas,
versions, counts, and non-secret routing metadata only when the question
requires them.

## Mutation Requests

This reference does not authorize mutation requests. For setup, install,
update, repair, capability policy, model-chain, category routing, provider, or
price changes, identify the appropriate specialized public workflow and stop
before writing. A separately authorized mutation begins a different workflow
with its own preview and verification contract.
rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file).Stars: 198kUpdated Claude CodeCodexUnverifiedInstall
doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.Stars: 198kUpdated Claude CodeCodexVerifiedInstall