Data Model: Skill Scanner Validation Errors
Skill Submission
Represents a skill package submitted for scanning.
Fields:
archive: Uploaded package containing one or more files.skill_directory: Directory selected from the uploaded package that containsSKILL.md.skill_manifest: Metadata parsed fromSKILL.md, including required fields such asname.scan_options: Selected scan policy and analyzer options.
Validation Rules:
- A submitted package must contain at least one
SKILL.md. - The selected skill definition must include all required metadata fields.
- Validation errors caused by malformed skill content are caller-correctable and must not be reported as internal service faults.
Validation Failure
Represents a caller-correctable problem detected before or during skill loading.
Fields:
category: Stable failure class, expected to be validation-related.message: Actionable, sanitized validation detail such as a missing required field.status: Validation response category that automated callers can distinguish from internal service faults.
Validation Rules:
- Message must not include stack traces, credentials, tokens, or unnecessary runtime filesystem paths.
- Message should preserve the actionable cause provided by the scanner loader when safe to expose.
- Failure category must be distinct from internal service failure.
Scan Result
Represents the outcome returned to a scanner caller.
Fields:
scan_id: Unique scan identifier for successful scans.skill_name: Name resolved from the submitted skill for successful scans.is_safe: Safety verdict for successful scans.max_severity: Highest finding severity for successful scans.findings_count: Count of findings for successful scans.findings: Structured findings for successful scans.validation_error: Validation failure detail for malformed submissions.
State Transitions:
Submitted
|-- Validated -> Scanned -> ScanResult(success)
|-- Invalid -> ValidationFailure
`-- UnexpectedFault -> InternalFailure
Relationships:
- A
Skill Submissionproduces exactly one outcome: successfulScan Result,Validation Failure, or internal failure. - A
Validation Failureis derived from a submitted skill's content, not from platform storage.
Storage Impact
No persisted storage changes are required. This data model documents API outcome semantics only.