Receipt Subscription Cleaner
Analysis
This read-only receipt-audit skill is coherent and includes privacy guardrails, but users should remember it may handle sensitive receipt or email data.
Findings (3)
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.
Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.
If API access is explicitly requested, use read-only scopes only.
The skill may use delegated email or API access when the user explicitly asks, which is sensitive account authority even though it is purpose-aligned and limited to read-only scopes.
Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.
Receipt sources (email export, PDF folder, or CSV list).
The main inputs can contain sensitive financial, address, and purchase-history information; the skill acknowledges this with local-storage and redaction guidance.
Optional: trigger when a new receipt arrives. Payload should include sender, subject, date, and attachment reference.
Optional webhook-style receipt triggers introduce an inbound data flow containing email metadata and attachment references; the artifact also says to treat inbound data as untrusted.
