Supermemory Search
Search past coding sessions and decisions to recall what you worked on before.
The skill only allows Bash(node:*) tool, which is a limited permission, no malicious behavior found. API key is passed via environment variable and set by user, data flow relies on external service supermemory.ai, no explicit data confidentiality or rollback mechanism. No user confirmation, automatic search may trigger unintentionally. Deducted heavily but no red line.
Skill provides clear CLI invocation but key dependencies (script, API service) not verified. Static review cannot confirm actual execution, no error handling or failure feedback described. Low reliability score.
Skill targets clear scenario (retrieve past session info) and trigger condition (user asks about past work), but doesn't specify non-fit cases like no network or missing API config. Environment fit: depends on overseas service supermemory.ai, may be unreachable in mainland China, reducing adaptability.
Docs well-structured with install, config, examples, migration, but lacks version history, license info in SKILL.md (only README mentions MIT), maintenance responsibility unclear.
Static review cannot verify actual effect, but description plausible, output is formatted memory results requiring further user review. Direct usability not confirmed, limited score.
Repo has CI workflows and test script (package.json test command), but test file not provided, cannot independently verify. Key claims like memory recall lack third-party evidence.
- Core functionality depends on overseas service supermemory.ai, may be unreachable in mainland China networks; consider alternatives or note in docs.
- Skill allows automatic memory search which may trigger unintentionally, involving privacy data; recommend adding user confirmation.
- API key management relies on user-set env var; remind users to store securely to avoid leak.
- Lacks clear rollback and data deletion details; user data may be stored by external service; review privacy policy.
What does this skill do, and when should you use it?
Supermemory Search is one of two skills bundled in the claude-supermemory plugin. It enables Claude Code to search a persistent memory store (Supermemory) for past coding sessions, decisions, and saved information. The skill runs a Node.js script that accepts a user query and an optional scope flag (--user, --repo, or --both) to retrieve relevant memories, presenting them with timestamps and relevance scores. It helps developers recall previous work, implementation details, or coding preferences, improving continuity across sessions. This skill is search-only; saving is handled by the sibling skill supermemory-save. It requires a Supermemory Pro account and an API key.
Executes node "${CLAUDE_PLUGIN_ROOT}/scripts/search-memory.cjs" with the user's query and optional scope flags (--user, --repo, --both) to query Supermemory for matching memories. The script returns formatted results with timestamps and relevance scores, which are presented to the user.
- A developer wants to recall what they worked on yesterday and asks 'What did I do yesterday?'.
- A developer needs to remember how authentication was implemented in a project, using the --repo scope.
- A developer wants to know their own coding preferences, using the --user scope.
- A developer returns to a project after a break and wants to quickly catch up on team progress by searching project memories.
What are this skill's strengths and limitations?
- Cross-session and team-shared memory retrieval.
- Flexible scoping options (personal, repo, both).
- Seamless integration with Claude Code, no manual invocation.
- Results include timestamps and relevance scores for easy filtering.
- Requires a paid Supermemory subscription and API key.
- Only covers search; saving functionality is in a separate skill.
- Needs Node.js 18+ environment.
- No evidence of a test suite; reliability unverified.
How do you install this skill?
Install via Claude Code plugin marketplace: /plugin marketplace add supermemoryai/claude-supermemory then /plugin install supermemory. Requires Node.js 18+ on PATH. Set the environment variable SUPERMEMORY_CC_API_KEY (obtained from app.supermemory.ai). Optional global settings file at ~/.supermemory-claude/settings.json and per-project config at .claude/.supermemory-claude/config.json can be used. License is MIT.
How do you use this skill?
Simply ask about past work or previous sessions, e.g., 'How did we implement auth?' or 'What did I work on yesterday?'. The skill will automatically run the search script; you can also explicitly specify scope by using natural language or rely on the default --both scope.