self-improving-agent-ecc
What this skill does
Self-improving agent system that captures atomic learnings (instincts) with confidence scoring, supports project-scoped vs global scope, and evolves instincts into skills/commands/agents.
github/leoyeai - Data Exfiltration - 2.1k stars
Threat analysis
Skill info
pkg:github/LeoYeAI/openclaw-master-skills@e5199b5?skill=self-improving-agent-eccAssessments (3)
Data Exfiltration
Data Exfiltration via local-llm-review
hooks/observe.sh
The script contains logic to detect projects and store observations in $HOMUNCULUS_DIR/observations.jsonl. While not explicitly exfiltrating data, the potential for this script to collect and store seCommand Injection Risk
Command Injection Risk via local-llm-review
scripts/error-detector.sh
The script checks for errors in CLAUDE_TOOL_OUTPUT and outputs a <error-detected> reminder. If the error detection logic is not properly sanitized, it could be exploited to inject malicious commands oPersistence Risk
Persistence Risk via local-llm-review
scripts/activator.sh
The script outputs a <self-improvement-reminder> that encourages logging to .learnings/ using the self-improvement skill format. This could be used to persist knowledge across sessions, potentially caBadge
Add the Anomity scan badge for self-improving-agent-ecc to your README.
How Anomity governs this at runtime
Scan-time vetting tells you what a skill says it will do. Anomity's Endpoint Sensor sees what agents actually do: it discovers skills alongside every other AI artifact on the endpoint, and runtime governance can allow, deny, or log the tool calls a skill triggers. Policy violations route to your SIEM, Slack, email, or Jira, backed by a queryable 90-day audit trail.
Book a 30-minute demo to see your own skill inventory.
Methodology and disputes
Every skill is assessed by the Anomity Skill Intelligence engine against its public source; findings indicate risk patterns, not confirmed exploitation. Maintainer of self-improving-agent-ecc? Report an issue or request a rescan.
