Caveat verdict
execution-harness-hub
execution-harness
This is a pure documentation/navigation hub for agent reliability patterns, with no code execution, data exfiltration, or malicious behavior present.
⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.
Automated static analysis — not a human review. Caveat flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.
Permission integrity
agent_memory
Findings (12)
Recursive delete from root or home — destructive command
tool-governance/references/graduated-permissions.md · code · rm -rf /
Pipe to sh — executes piped content as shell commands
tool-governance/references/tool-input-guard.md · code · |sh
Pipe to bash — executes piped content as shell commands
error-recovery/references/graceful-degradation.md · code · | Bash
Accesses sensitive system files
tool-governance/references/tool-input-guard.md · code · /etc/passwd
Pipe-to-shell pattern (curl | sh) — supply chain attack vector
docs/ata-execution-harness-v4.md · prose · downgraded · curl | sh
References sudo — requests elevated privileges
tool-governance/references/graduated-permissions.md · code · sudo
subprocess execution — runs system commands from Python
context-memory/tests/test_compaction_extract.py · prose · downgraded · subprocess.run(
Instructs covert action — may act without user awareness
principles.md · prose · downgraded · Silently
Sets world-executable permissions
tool-governance/references/graduated-permissions.md · code · chmod 777
apt-get install — installs system packages
tool-governance/SKILL.md · prose · downgraded · apt-get install
References agent configuration files
context-memory/references/extended-patterns.md · prose · downgraded · CLAUDE.md
References agent memory files
context-memory/SKILL.md · prose · downgraded · memory.md
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class A/D). Final tier capped at Caution — cannot be lifted by any downgrade, example-payload opt-in, or allowlist.
Permissions & capabilities
No declared permissions — minimal attack surface.
agent_memory Is this flag fair?
Thanks — recorded.