memory-compression-system
v3.0.1Integrated memory management and extreme context compression for OpenClaw. Combines memory management, compression, search, and automation in one unified skill.
⭐ 0· 801·6 current·6 all-time
by@hassffw
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The name/description (memory compression + scheduling) matches what the included scripts do: compress, backup, cleanup, create cron jobs, and update a local search index. Creating backups of workspace memory and adding OpenClaw cron jobs is coherent with the stated purpose. However the SKILL.md advertises many management/search utilities (decompress.sh, search.sh, list.sh, backup.sh, restore.sh, logs.sh, metrics.sh, search-history.sh, etc.) that are referenced but not present or are only stubbed—so the claimed feature set is larger than the delivered files.
Instruction Scope
Scripts (compress.sh, install.sh, cleanup.sh, enable/disable/status) read /home/node/.openclaw/workspace/memory/*, create local backups (tar.gz), create/update cron jobs via the openclaw CLI, and write logs and index files under the skill directory. That behavior is expected for a compression/backup skill but is sensitive because it collects and archives memory files from the agent workspace. Several SKILL.md instructions call scripts that are missing or placeholders (e.g., health.sh is a stub that exits), so the instructions are inaccurate and grant broad runtime expectations that the package does not actually implement.
Install Mechanism
There is no remote installer or download-from-URL step; installation is via the provided scripts (install.sh). No external archives or third-party packages are fetched by URL. install.sh writes configuration and creates directories inside the OpenClaw workspace. This is lower-risk than arbitrary network downloads, but it still writes files and enables scheduled tasks locally.
Credentials
The skill requests no external credentials or environment variables beyond optional debug/testing flags. That is proportionate. However the scripts operate on local workspace data (reading memory files and creating backups), which is sensitive: absence of requested credentials is appropriate, but the actions access agent memory and produce persistent backups and logs—ensure you are comfortable with local data collection.
Persistence & Privilege
always is false, but install/enable scripts attempt to create an OpenClaw cron job that will cause periodic autonomous runs of the compression script (every 6 hours). Autonomous scheduling aligns with the skill's purpose but raises operational risk because it gives the skill recurring execution without manual invocation. The skill does not request system-level privileges beyond workspace writes, and it avoids modifying other skills' configurations.
What to consider before installing
Read and test before enabling. Specific points to check:
- Missing/placeholder scripts: SKILL.md references many helper scripts (search.sh, decompress.sh, backup.sh, restore.sh, logs.sh, metrics.sh, etc.) that are not present or are stubs (health.sh). Expect the advertised feature set to be incomplete.
- Local data access: compress.sh and install.sh read /home/node/.openclaw/workspace/memory/*.md, create tar.gz backups, and write compressed files/logs under the skill directory. If those memory files contain sensitive data you do not want archived, do not enable the skill or turn off backups (BACKUP_BEFORE_COMPRESSION=false) and review which paths are processed.
- Scheduled autonomous runs: enable.sh / install.sh try to add an OpenClaw cron job that will run the compression periodically. That gives the skill recurring, automated execution—review the cron job payload and schedule before creation (or skip cron creation and run manually until you audit the scripts).
- Test-mode and dry-run: use the --test / --dry-run flags and inspect logs in a safe environment first. Run install.sh and install in an isolated or non-production workspace to confirm behaviour.
- Code quality issues: some script logic appears brittle or buggy (jq usage, some arithmetic expressions), so expect possible runtime errors. Inspect scripts for any commands you don't want run and ensure required utilities (tar, gzip, base64, crc32, jq, uuidgen) are present or disabled.
- If you do not trust the author or cannot fully audit the scripts, avoid enabling automatic cron scheduling and run only manual compressions in an isolated environment.Like a lobster shell, security has layers — review code before you run it.
latestvk97463phxxd9paxfdpf0daxdw1816x47
License
MIT-0
Free to use, modify, and redistribute. No attribution required.