agentmemory-hooks
rohitg00/agentmemory
Automatic memory capture hooks across agent sessions—zero-LLM observation recording.
What is agentmemory-hooks?
The agentmemory plugin registers lifecycle hooks that capture observations automatically without manual saves. Use this to understand how memory is recorded during tool use, prompts, and session boundaries, or to debug missing observations and tune capture behavior.
- Captures session start and end to frame units of work for handoff resumption
- Records tool-use changes and intent via hooks on execution and prompt submission
- Preserves context before host trimming with pre-compact hooks
- Links commits to sessions for commit-context and commit-history features
- Operates zero-LLM by default; token-spending summarization and injection are optional
How to install agentmemory-hooks
npx skills add https://github.com/rohitg00/agentmemory --skill agentmemory-hooks- agentmemory plugin installed and enabled
- agentmemory server running (default http://localhost:3113)
How to use agentmemory-hooks
- 1.Install the plugin via /plugin marketplace add rohitg00/agentmemory and /plugin install agentmemory
- 2.Confirm the server is running at http://localhost:3113 to see observations land live
- 3.Verify plugin is enabled if observations are missing; check TROUBLESHOOTING.md if needed
- 4.Optionally enable AGENTMEMORY_AUTO_COMPRESS to summarize observations with LLM
- 5.Optionally enable AGENTMEMORY_INJECT_CONTEXT to inject summaries back into agent context
Use cases
- Debugging why observations are missing from memory during a session
- Understanding what gets automatically recorded without calling memory_save
- Tuning capture behavior by enabling or disabling auto-compress and context injection
- Tracing tool execution and prompt intent through the observation pipeline
- Linking code commits to agent sessions for context-aware history
- Developers building with Claude Code or Cursor agents
- Engineers debugging agent memory and observation gaps
- Teams tuning memory capture to balance coverage and token cost
agentmemory-hooks FAQ
No. The hooks capture observations automatically during tool use, prompts, and session boundaries. memory_save is not required for routine capture.
Capture itself is zero-LLM. Token-spending features (auto-compress and context injection) are separate opt-ins controlled by AGENTMEMORY_AUTO_COMPRESS and AGENTMEMORY_INJECT_CONTEXT.
Confirm the plugin is enabled and the server is running at http://localhost:3113. See TROUBLESHOOTING.md for detailed troubleshooting steps.
Session start/end, tool-use, prompt-submit, pre-compact, and post-commit hooks. Full details are in REFERENCE.md generated from plugin/hooks/hooks.json.
A post-commit hook links commits to sessions, enabling commit-context and commit-history features to provide context-aware code history.
Full instructions (SKILL.md)
Source of truth, from rohitg00/agentmemory.
name: agentmemory-hooks description: The agentmemory plugin hooks that capture observations automatically across the agent session lifecycle. Use when explaining how memory gets captured without manual saves, when debugging missing observations, or when tuning what gets recorded. user-invocable: false
The Claude Code plugin registers lifecycle hooks so memory is captured automatically. You do not have to call memory_save for routine work; the hooks observe tool use, prompts, and session boundaries and write observations for you.
Quick start
Install the plugin and the hooks register themselves:
/plugin marketplace add rohitg00/agentmemory
/plugin install agentmemory
Watch observations land live at http://localhost:3113.
What the hooks do
- Session start and end frame each unit of work and let
handoffresume it. - Tool-use hooks capture what changed and why, the raw material for
recallandrecap. - Prompt-submit captures intent. Pre-compact preserves context before the host trims it.
- A post-commit hook links commits to sessions, which powers
commit-contextandcommit-history.
Important
- Capture is on by default and is zero-LLM. Turning observations into LLM summaries (
AGENTMEMORY_AUTO_COMPRESS) and injecting them back into context (AGENTMEMORY_INJECT_CONTEXT) are separate opt-ins because they spend tokens. - If observations are missing, confirm the plugin is enabled and the server is running. See ../_shared/TROUBLESHOOTING.md.
See also
- agentmemory-config for the capture and injection flags.
- The handoff, recap, and session-history skills consume what these hooks record.
Reference
The exact registered hook events live in REFERENCE.md, generated from plugin/hooks/hooks.json.
Related skills
More from rohitg00/agentmemory and the wider catalog.
agentmemory-mcp-tools
Index of agentmemory MCP tools: which tool to call, exact parameters, and when to use each.
agentmemory-rest-api
HTTP REST API for agentmemory memory server—save and recall agent knowledge over HTTP.
commit-context
Trace code back to the agent session that created it—answer 'why is this here' and 'what was the agent doing'.
commit-history
List recent git commits linked to agent sessions, filtered by branch or repo.
forget
Delete specific observations from agentmemory with explicit user confirmation.
handoff
Resume the most recent agent session for your project, picking up from any unanswered question.