self-improving-agent-1-0-0

clawhub:self-improving-agent-1-0-0

View source

70/100

First Seen

Feb 18, 2026

Last Scanned

Feb 22, 2026

Findings

Score

70/100

HIGH 2

Findings (2)

HIGH

Self-modifying agent instructions PROMPT_INJECTION_016

prompt-injection L21

Detects skills that write or promote content into agent instruction files

Promote to `CLAUDE.md

FIX

Remove or restrict the skill's ability to modify CLAUDE.md or agent configuration files. Self-modifying agent configurations can be exploited for persistent attacks.

FP?

Likely FP if the skill is a project management tool where updating CLAUDE.md is an intended workflow feature (e.g., appending project notes), though the risk remains.

HIGH

Self-modifying agent instructions PROMPT_INJECTION_016

prompt-injection L217

Detects skills that write or promote content into agent instruction files

promote to CLAUDE.md

FIX

Remove or restrict the skill's ability to modify CLAUDE.md or agent configuration files. Self-modifying agent configurations can be exploited for persistent attacks.

FP?

Likely FP if the skill is a project management tool where updating CLAUDE.md is an intended workflow feature (e.g., appending project notes), though the risk remains.