What does this skill do, and when should you use it?
The supermemory-save skill enables Claude Code to save user-specified project knowledge—such as architectural decisions, significant bug fixes, design patterns, or critical implementation details—into Supermemory for persistent team memory. It works by analyzing the conversation, formatting content into a structured block, and invoking a Node.js script to store it. The skill requires a Supermemory Pro subscription, a valid API key, and Node.js 18+. It ensures team knowledge is retained across sessions and projects, enhancing collaboration and continuity.
It reads the current conversation to understand the user's save request, formats the content into a [SAVE] block with username, date, user's goal/problem, Claude's suggestion, decision made, and key details, then runs the Node.js script ${CLAUDE_PLUGIN_ROOT}/scripts/save-project-memory.cjs to save it to Supermemory as team memory.
- A developer fixes a complex bug and asks to record the fix and involved files for team reference.
- An architect decides on a new design pattern and wants to document it for the team.
- A project manager requests to save important implementation details for future developers.
- A team member wants others to know about a critical technical decision and its context.
How do you install this skill?
- The skill depends on overseas service supermemory.ai, may be unreachable from mainland China; consider alternatives or add network configuration notes.
- Skill allows Bash(node *) without user confirmation; recommend explicit prompt before saving and disclose data sent to external service.
- Lacks error handling and failure feedback; consider adding handling for abnormal inputs.
- Shell / CLI
- Network access
- Local filesystem
Node.js 18+Supermemory account
- Ensure Node.js 18+ is installed and on PATH. 2. In Claude Code, run
/plugin marketplace add supermemoryai/claude-supermemoryand then/plugin install supermemory. 3. Set your API key:export SUPERMEMORY_CC_API_KEY="sm_..."(get one at app.supermemory.ai).
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/supermemoryai/claude-supermemory.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/plugin/skills/supermemory-save" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Please save our decision about X.
Naturally ask Claude to save something, e.g., "Please save our decision about X." Claude will analyze the conversation, format the content, and run the save script. Ensure the plugin is installed and the API key is set.
What are this skill's strengths and limitations?
- Persists team knowledge across sessions and projects
- Automates content formatting, saving time
- Deeply integrated with Claude Code, no manual steps
- Requires Supermemory Pro or above, potentially costly
- Relies on external service (Supermemory API)
- Authentication via API key is somewhat involved
- No built-in test coverage, script may be fragile
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Supermemory Save Skill this page | 34 · Not recommended | ★ 2.8k | 4d ago | — |
| Supermemory Search | 37 · Not recommended | ★ 2.8k | 4d ago | — |
| Compound Learning Capture | 53 · Use with care | ★ 25k | 3d ago | MIT |
| Compound Refresh — Learning Docs Maintenance | 52 · Use with care | ★ 25k | 3d ago | MIT |
| ADO Wiki Converter ✓ Microsoft · Official | 54 · Use with care | ★ 2.7k | 3mo ago | MIT |
How did FollowSkills review this skill?
Evidence shows the skill allows Bash(node *) to execute external script, but no user confirmation mechanism, data flow transparency, sensitive-data handling, or rollback is mentioned. API key is provided via env var, not directly handled by skill. There is potential to write to external service without clear disclosure. Thus trust is low, deduct mainly for broad permissions, lack of confirmation and transparency.
Skill description provides format template and command to invoke script, but no error handling, abnormal input feedback, or test evidence. CI workflow only covers lint and build, not tests. Static review cannot execute, and lack of test suite limits reliability score. Main deductions for lack of robustness and failure feedback.
Skill clearly for saving important project knowledge, scenario clear, but no explicit trigger boundaries (when not to trigger) and depends on external overseas service Supermemory, may be unreachable from mainland China, so environment fit limited. Score moderate low.
README provides installation, config, commands, privacy policy, and MIT license, fairly complete. But lacks version history and changelog, maintenance responsibility unclear. Overall readable but governance incomplete.
Skill describes process to save knowledge, but output direct usability not verified. Depends on external service, no example outputs or comparative advantage. Static review cannot confirm actual completion, so score low.
CI workflow and package.json exist, but CI does not run tests, and no test cases covering key paths. Thus verifiability low, only author claims and build steps.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →