Reflect Notes

**Judgement:** A solid, focused integration for users already committed to Reflect. It’s best viewed as a one-way capture pipe from an OpenClaw agent into a secure personal knowledge graph, not as a full note-management interface. **Strengths & Fit** - Works well if you: - Already use Reflect as your primary notes/PKM system. - Want the agent to *capture* todos, notes, and links without giving it broad read access to your existing notes. - Prefer privacy-preserving behavior: the API is effectively write-only for note content. - The primitives (append daily note, create note, create/save links) are exactly what you need for turning conversational output into persistent artifacts. **Key Limitations / Risks** - **No read/edit/search of notes:** The agent cannot inspect, edit, or delete note bodies due to E2E encryption and append-only semantics. It also can’t search your graph. This severely limits any workflow that needs context from previous notes or cleanup of mistakes. - **Potential clutter / duplication:** Because content is append-only, misconfigured prompts or overly verbose agents can spam daily notes or lists with redundant or low-quality entries you must manually clean up in Reflect. - **External dependency & tokens:** Requires managing Reflect OAuth token and graph ID as environment variables; token expiry or misconfiguration will break the integration. Security of the token is critical, since it allows writing into your graph. - **Partial retrieval:** Only links/books can be fetched via the provided endpoints; you can’t use the agent as a full Reflect client. **Recommended Scenarios** - Use this skill when you want: - **Automatic capture of outcomes** (todos, decisions, summaries) from chat into a trusted external system. - **Link saving and highlight capture** from URLs discussed in-session. - **Daily journaling / running log** where the agent appends short entries or bullets to your daily note. - Avoid relying on it for workflows that require: - Reading past notes for context-aware reasoning. - Editing or curating existing notes programmatically. - Complex knowledge management operations (search, re-structuring, bulk cleanup). Overall, this is a good, relatively safe choice for one-way logging into Reflect for users who already live in that ecosystem, but it’s not suitable as a general-purpose notes backend for sophisticated agent reasoning or note maintenance.