Memory Archiver

v1.0.0

定期扫描并压缩超过指定天数的内存日志,提炼关键信息更新长期记忆,释放存储空间并清理过期文件。

0· 94·1 current·1 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
medium confidence
Purpose & Capability
Name and description (archive old memory logs, extract key info, update MEMORY.md, compress and delete old files) match the actions described in SKILL.md (scan memory/*.md, extract, merge, zip, delete). No unrelated credentials, binaries, or installs are requested.
Instruction Scope
Instructions operate on local memory/ files (memory/YYYY-MM-DD.md, MEMORY.md, memory/archive/) which is appropriate, but they reference an undefined step/tool 'memory_search' and broadly instruct AI to 'identify important decisions/context/todos' which is open-ended. The flow deletes original files by default (delete_after_archive=true) without describing dry-run, confirmation, or backup — this is functionally coherent but a safety/usability concern.
Install Mechanism
No install spec and no code files (instruction-only). This is low-risk from an installation perspective — nothing is downloaded or written to disk by an installer.
Credentials
The skill requests no environment variables, credentials, or config paths. The requested scope (local memory directory operations) is proportional to the stated purpose.
Persistence & Privilege
always is false and autonomous invocation is default (normal for skills). The skill does perform file deletes and writes to MEMORY.md and archive/ within the user's workspace, which is appropriate for its purpose but requires the agent to have local file write/delete permission.
Assessment
This skill appears to do what it says: find old files under memory/, summarize them into MEMORY.md, compress them into memory/archive/, and (by default) delete the originals. Before installing or running it: 1) Confirm what 'memory_search' refers to (a built-in memory index, a CLI tool, or something else) so you know what data will be read. 2) Test with delete_after_archive=false or run in a dry-run mode on a copy so you don’t lose data unexpectedly. 3) Backup MEMORY.md and the memory/ folder before enabling automatic runs. 4) Verify the cron/openclaw commands referenced exist in your environment and that automated runs have the intended permissions. 5) If you’re concerned about data exfiltration, ensure the agent environment has no outbound integrations enabled or audit what external calls (if any) the platform makes when summarizing. If these checks look good, the skill is coherent and reasonable to use.

Like a lobster shell, security has layers — review code before you run it.

latestvk97dd1edg91jp0ft60q610k0rs83qv07

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments