ClawHub

Agent Stability Framework

v1.0.0

Provides a framework to prevent agent drift, catch faults, and maintain consistent on-character behavior across sessions and models.

⭐ 0· 404·1 current·1 all-time
bydp-del@donovanpankratz-del
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 and description match the actual contents: documentation, templates, and prompt fragments for preventing drift, catching faults, and enforcing a persona. There are no unrelated env vars, binaries, or install steps requested.
✓
Instruction Scope
SKILL.md and included templates instruct the user/agent to copy files into the workspace, add standing orders and a pre-send gate to the system prompt, create logs, and run audits — all consistent with a stability framework. The instructions do not ask the agent to read unrelated system files, access credentials, or transmit data to external endpoints.
✓
Install Mechanism
No install spec, no code files, and no downloads. This is instruction-only, so no code will be written or executed by an automated installer.
✓
Credentials
The skill requests no environment variables, credentials, or config paths. The templates reference only files the user is expected to create in their agent workspace, which is proportionate to the stated purpose.
ℹ
Persistence & Privilege
The skill does not set always:true and is user-invocable. However, its recommended actions (editing the system prompt, adding standing orders, and copying templates into the workspace) change agent behavior persistently — this is expected for a stability framework but is a material, lasting change the user should review and control.
Assessment
This skill is essentially a set of docs and prompt/templates to change how your agent behaves — there is no code or external installer. Before installing: (1) Review all template files to ensure the standing orders fit your safety needs (some rules like 'NEVER add disclaimers' or 'NEVER explain why unless asked' may remove helpful safety/clarity checks). (2) Back up your current system prompt and agent workspace. (3) Test the framework in an isolated or sandbox agent first to observe effects. (4) Keep monitoring logs and be ready to revert standing orders that cause unexpected behavior (loss of necessary warnings, refusal to perform required explanations, or interference with safety constraints). If you want an automated installer or integration with external systems, request more details — currently this skill only provides manual instructions and templates.

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

latestvk974hpfxaqtmmtz4fhas531smh81t8h4

License

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

Comments