Escalate
v1.0.2Self-learn to decide when to act, when to ask, and which actions should always need approval.
⭐ 5· 1.8k·16 current·17 all-time
byIván@ivangdavila
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (decide when to act vs ask) matches the instructions: creating and reading local escalation memory, decisions log, and domain overrides is necessary for persistent ask-vs-act behavior. No unrelated binaries, credentials, or external services are requested.
Instruction Scope
Runtime instructions read and write only local files in ~/escalate and may read workspace steering files (AGENTS and SOUL) to align behavior. This is coherent with the skill's purpose but worth noting: the skill will read those workspace steering files and suggests adding small snippets to them — it documents that it will show snippets and wait for explicit approval before writing.
Install Mechanism
Instruction-only (no install spec, no code to download). Low risk: nothing is written to disk by an installer beyond the documented local state the skill itself creates under user control.
Credentials
No environment variables, credentials, or config paths beyond the declared ~/escalate location are requested. The local storage and file permissions (chmod 700/600) are proportionate to preserving private escalation memory.
Persistence & Privilege
The skill persists state under ~/escalate (intended). always:false and no autonomous install behavior reduce risk. Note that persistent logs and rules will live on disk across sessions — expected for its purpose but something the user should understand and manage.
Assessment
This skill is internally coherent and appears to do what it says: it keeps local escalation policies in ~/escalate, reads workspace steering files to decide when to act vs ask, and asks for explicit approval before making workspace edits. Before installing, confirm you are comfortable with: (1) persistent local storage under ~/escalate (review and remove sensitive entries if any), (2) the skill reading your AGENTS/SOUL steering files, and (3) giving permission before it writes snippets into workspace files. Also verify the skill's origin/homepage if you require stronger provenance guarantees.Like a lobster shell, security has layers — review code before you run it.
latestvk976r7jh3hr85c6p9g18n1591182kqax
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
🚨 Clawdis
OSLinux · macOS · Windows
