OpenClaw Memory Curator
v1.0.0Organize, deduplicate, summarize, and compress OpenClaw or Clawd memory files. Use when the user asks to organize memory, reduce duplication, convert raw dai...
⭐ 0· 52·0 current·0 all-time
byzeruel@1217047020
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description match the included files and behavior. The SKILL.md and README describe organizing, deduplicating, summarizing, and backing up memory files, and the included script implements report/backup/validate operations for the workspace. No unrelated binaries, env vars, or config paths are requested.
Instruction Scope
Instructions ask the agent to read AGENTS.md, MEMORY.md, and memory/*.md and to back up before making edits — all appropriate for curating memory. One minor mismatch: SKILL.md documents a 'rewrite' workflow (editing/compressing notes), but the bundled script only implements reporting, backup, and validation (no automated rewrite). That means actual rewriting would be performed by the agent itself (file edits), which the SKILL.md correctly constrains (require user approval and backups) but is not enforced by code.
Install Mechanism
This is an instruction-only skill with a small helper script included; there is no install spec, no downloads, and no execution of external installers. Low installation risk.
Credentials
The skill requires no environment variables, no credentials, and no config paths. Its file I/O is limited to the declared workspace structure (AGENTS.md, MEMORY.md, memory/, memory-backups/).
Persistence & Privilege
always:false and normal autonomous invocation defaults. The skill does not request persistent platform-level privileges or modify other skills. Note that because the script doesn't perform rewrites, any file edits will be performed by the agent itself when invoked — SKILL.md instructs to get explicit user approval before edits.
Assessment
This skill appears coherent and low-risk: it reads and backs up local workspace files and contains no network or credential usage. Before installing or allowing it to run: (1) verify the workspace path it will act on (default examples use /root/clawd), (2) review AGENTS.md and the repository files (scripts/curate_memory.py) yourself so you know exactly what will be read and backed up, (3) if you allow the agent to perform edits, require the confirmation step the SKILL.md specifies (it relies on the agent to ask for approval), and (4) keep regular backups of your workspace. If you want automated rewriting rather than manual edits, note the included script does not implement that — the agent will perform file edits if instructed.Like a lobster shell, security has layers — review code before you run it.
latestvk97904emha5jhnrspxvjyemq9983z4bc
License
MIT-0
Free to use, modify, and redistribute. No attribution required.