hermes-context-management

Decides where to persist declarative knowledge across Hermes context mechanisms.

Updated Jul 10, 2026
One-click install
npx skills add https://github.com/timchap/dot-hermes --skill hermes-context-management-timchap
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hermes-context-management
Source: https://github.com/timchap/dot-hermes/tree/main/skills/meta/hermes-context-management
Command: npx skills add https://github.com/timchap/dot-hermes --skill hermes-context-management-timchap

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Hermes offers many context-loading mechanisms (SOUL.md, .hermes.md, AGENTS.md, memory, skills, references, templates, scripts), and misplacing knowledge wastes context space or leaks it into the wrong sessions. This Skill provides a decision framework for choosing the correct storage location based on scope, lifetime, and content type. ## Core Features & Use Cases - Placement Decision Framework: A step-by-step decision process mapping content types (identity, project rules, facts, procedures, external docs, templates, scripts) to the right Hermes storage mechanism. - Skill Library Quality Review: A systematic audit process covering inventory, duplication detection, frontmatter standards, and surgical patch-based fixes. - Pitfall Catalog: Documented failure modes such as ambiguous skill name collisions and Hindsight backend connectivity failures, with concrete fixes. - Use Case: After learning a new user preference or discovering a reusable workflow, run the decision process to determine whether it belongs in memory, a skill, or SOUL.md, then verify with the checklist. ## Quick Start Ask the agent to decide where a newly learned fact or procedure should be persisted in Hermes and have it walk through the placement decision process.

Frequently Asked Questions about hermes-context-management

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I decide where to store knowledge in Hermes?▼

Use the decision process: identity and mandates go to SOUL.md, project-specific rules to .hermes.md or AGENTS.md, persistent user or environment facts to memory, repeatable procedures to skills, external docs to references, boilerplate to templates, and deterministic actions to scripts.

What is the difference between SOUL.md and memory in Hermes?▼

SOUL.md holds identity and behavioral mandates loaded every session with a roughly 5k character budget. Memory stores compact declarative facts about the user or environment injected every turn, and should never contain task progress or soon-stale information.

When should I use .hermes.md versus AGENTS.md?▼

Use .hermes.md for Hermes-specific project rules that inherit hierarchically up to the git root. Use AGENTS.md or CLAUDE.md when other agents like Claude Code, Codex, or OpenCode will also work in the repository.

Why does hindsight_retain fail with HTTP 500?▼

Hindsight uses a local LLM for fact extraction, and if the configured Lemonade model is removed or renamed, retain calls fail with HTTP 500 model_not_found. Always run hindsight_recall("any") first to verify connectivity before bulk stores.

How do I fix an ambiguous skill name error in Hermes?▼

The error occurs when a flat file like foo.md and a directory foo/SKILL.md share a name. Read both candidates, merge unique content into the directory-form SKILL.md, delete the stray flat file, then re-run skill_view to confirm a single match.