ClawHub

Prompt Diet

v1.0.0

Analyze and optimize OpenClaw system prompt token usage by compressing workspace files (AGENTS.md, SOUL.md, USER.md, TOOLS.md, IDENTITY.md, HEARTBEAT.md, MEM...

⭐ 0· 31·0 current·0 all-time
by@dgkim311
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
Name/description match actual behavior: it measures token usage and recommends/edits workspace .md files. No unrelated binaries, env vars, or network access are requested.
✓
Instruction Scope
SKILL.md limits operations to workspace markdown files, requires explicit user approval before applying 🟡 review edits, and outlines safe/skip rules. It does read and may modify files in the workspace (expected for this task); it does not instruct exfiltration or access to unrelated system areas.
✓
Install Mechanism
Instruction-only skill with a small included Python script. No install spec, no external downloads, and optional dependency on tiktoken (falls back to chars/4 approximation).
✓
Credentials
No required environment variables, credentials, or config paths are declared. The operations (reading/writing workspace files) are proportional to the stated goal.
✓
Persistence & Privilege
always is false and the skill does not auto-register periodic jobs. It edits only workspace files (with backups) and does not claim to modify other skills or system-wide agent settings.
Assessment
This skill appears coherent and safe for its stated purpose, but you should: (1) confirm the workspace path before running (it defaults to /home/aif/.openclaw/workspace/ or a local .openclaw/workspace), (2) review recommendations and approve any 🟡 items before edits, (3) ensure sensitive tool auth or secret references in TOOLS.md or MEMORY.md are not accidentally included in automatic changes (the rules say to skip them but double-check), and (4) keep the backups the skill creates until you verify behavior. Optionally install the tiktoken package for more accurate token counts (pip install tiktoken).

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

latestvk979f1c3vvy12bpp8qjza7dbr184hd8r

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Comments