Revenue Recovery

AdvisoryAudited by Static analysis on May 1, 2026.

Overview

No suspicious patterns detected.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

If the agent runs onboarding with a malformed or attacker-controlled client ID, it could write outside the intended client folder or run unintended local Python code.

Why it was flagged

The user-provided client ID is used directly in a filesystem path and shell-expanded into Python source without validation or escaping.

Skill content
CLIENT_ID="${1:?client_id required}" ... ROOT="$HOME/.openclaw/workspace/data/xzenia/clients/$CLIENT_ID" ... cfg=os.path.expanduser('$ROOT/client_config.json')
Recommendation

Validate client IDs with a strict allowlist such as letters, numbers, dash, and underscore; resolve paths under a fixed base directory; and pass values to Python through argv or environment variables rather than expanding them into source code.

What this means

Running setup may install different dependency versions over time, which can affect reproducibility and supply-chain assurance.

Why it was flagged

The setup script downloads third-party packages at run time using broad lower-bound version ranges and no lockfile or hashes.

Skill content
pip install --quiet --upgrade pip || true ... pandas>=2.0 ... requests>=2.31 ... pip install --quiet -r /tmp/xzenia-requirements.txt || true
Recommendation

Pin dependency versions, use a lockfile or hashes, and review dependencies before running setup in a sensitive workspace.

What this means

Sensitive business or client data may remain on disk and influence later analyses.

Why it was flagged

The skill persists client financial events, contract terms, leakage findings, and recovery actions in a local SQLite database for reuse.

Skill content
DB_PATH="$ROOT/causal_memory.db" ... CREATE TABLE IF NOT EXISTS financial_events ... contract_terms ... leakage_findings ... recovery_actions
Recommendation

Use only approved data, segregate client workspaces, restrict filesystem access, and define cleanup or retention procedures.

What this means

Local files outside this skill may influence agent behavior and priorities if they are stale, modified, or untrusted.

Why it was flagged

The skill makes local project documents and a vague founder-priority rule part of its runtime guidance, even though those documents are not included in the reviewed artifacts.

Skill content
This skill operates under the following bound artifacts, in precedence order: ... projects/xzenia/docs/... ... Founder vector outranks local optimization.
Recommendation

Review the referenced governance files and ensure they cannot override explicit user instructions, approval requirements, or safety boundaries.