Back to skill
Skillv1.0.0
ClawScan security
Text To Video By Canva · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
SuspiciousApr 24, 2026, 6:24 PM
- Verdict
- suspicious
- Confidence
- medium
- Model
- gpt-5-mini
- Summary
- The skill appears to implement text→video by calling nemovideo.ai, but its name, declared requirements, and metadata contain inconsistencies and provenance gaps; it will upload user files and may leak environment/installation information, so review before installing.
- Guidance
- This skill will send your text and any uploaded files to a third-party backend (mega-api-prod.nemovideo.ai). Before installing: 1) Verify the publisher/source — the skill claims 'By Canva' but the API host is 'nemovideo.ai' and no homepage/source is provided. 2) Do not upload sensitive or private files unless you trust the service and its privacy terms. 3) Prefer creating a scoped, low-privilege NEMO_TOKEN (or review what the anonymous token grants) rather than supplying a long-lived credential. 4) Ask the publisher why the registry lists NEMO_TOKEN as required when the SKILL.md describes creating an anonymous token and why a config path (~/.config/nemovideo/) appears in metadata. 5) Consider refusing installation if provenance cannot be confirmed or if you cannot accept that the agent will transmit files and environment/installation metadata to the nemovideo.ai service.
Review Dimensions
- Purpose & Capability
- concernThe skill is named 'Text To Video By Canva' but all runtime endpoints point to mega-api-prod.nemovideo.ai and there is no homepage or source provenance. The declared primary credential (NEMO_TOKEN) is consistent with the API, but the mismatch between the Canva branding and the actual API host plus missing publisher information is a red flag for origin/legitimacy. The SKILL.md metadata also lists a config path (~/.config/nemovideo/) even though the registry shows 'required config paths: none' — this inconsistency suggests sloppy or conflicting configuration.
- Instruction Scope
- concernInstructions will upload user-provided files (multipart upload or URL) and create sessions/tokens on a third-party backend; they instruct the agent to check NEMO_TOKEN and, if missing, obtain an anonymous token via an API call. The skill also specifies headers derived from the agent's install path (X-Skill-Platform), which could reveal environment/installation details. Overall the actions are consistent with a cloud-render service, but the instructions permit automatic token creation and reference local config paths — both broaden the agent's runtime behavior beyond a simple conversion tool.
- Install Mechanism
- okNo install specification or code files are present; this is instruction-only, which reduces disk-write risk. There is no download/extract/install of third-party binaries described.
- Credentials
- concernOnly one environment variable (NEMO_TOKEN) is declared, which is appropriate for an API-backed service. However the SKILL.md states that if NEMO_TOKEN is missing the agent should obtain an 'anonymous' token by calling the API — yet the registry listed NEMO_TOKEN as required. That conflict (required vs optional) is incoherent. The metadata's config path (~/.config/nemovideo/) is also present in the SKILL.md, which could result in reading local configuration/config files not declared in the registry. The skill also requires inclusion of attribution headers that may leak install-path-derived info.
- Persistence & Privilege
- noteThe skill is not set to always:true and uses default autonomous invocation settings (normal). There is no install-time persistent agent modification described. However, the skill's header rules (X-Skill-Platform derived from install path) could cause the agent to expose its installation path; this is a privacy consideration rather than a privilege escalation.