Library
PassAudited by VirusTotal on May 12, 2026.
Overview
Type: OpenClaw Skill Name: sansfiction-library Version: 1.0.1 The skill is designed for managing a SansFiction library and scheduling reading reminders. It uses `curl` to interact with the `https://sansfiction.com/api/mcp` endpoint, authenticating with a `SANSFICTION_TOKEN`. The `skill.md` provides detailed instructions for the AI agent on how to handle user requests, manage the token securely (explicitly stating not to echo it), and schedule a daily reminder using OpenClaw's internal cron mechanism. All actions and instructions are aligned with the stated purpose, and there is no evidence of malicious intent, data exfiltration, unauthorized execution, or harmful prompt injection.
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.
Anyone with the token could potentially access or change your SansFiction library within that token’s permissions.
The skill needs an account token that can read and write SansFiction library data. This is expected for the stated purpose, but it is still sensitive account authority.
Requires a SansFiction personal token (read/write).
Use a dedicated SansFiction token, store it only in the intended OpenClaw config or secret mechanism, and revoke/regenerate it if it is exposed.
Pasting a token into chat or config can expose it if chat history, logs, or local config files are accessible to others.
The setup flow collects the token through chat and recommends saving it in local OpenClaw configuration. This is disclosed and tied to the integration, but users should treat the token as a secret.
Ask the user to paste the token once in this chat. Persist it (recommended): In `~/.openclaw/openclaw.json`...
Prefer a secure secret/config entry method when available, avoid sharing the token elsewhere, and ensure local OpenClaw config files are protected.
User requests can result in books being added, removed, marked finished, or progress being logged in the SansFiction account.
The skill can invoke SansFiction tools that mutate account library state. These actions are aligned with the skill’s library-management purpose and include ambiguity checks, but they are real account changes.
Use MCP search tools... Call the “add to library” tool... Call the “set status” tool... Call the “log progress” / “update progress” tool.
Review confirmations carefully, especially for remove or status-change requests, and ask the agent to show matches before acting when a title is unclear.
Your reading activity and library changes are sent to SansFiction as part of normal operation.
The skill communicates with an external MCP endpoint using the user’s bearer token. The endpoint is disclosed and purpose-aligned, but it transmits authenticated library queries and progress data to SansFiction.
Endpoint: `https://sansfiction.com/api/mcp` ... Use JSON-RPC with Bearer auth.
Use this only if you trust SansFiction with your library and reading-progress data, and verify the endpoint is exactly the documented SansFiction URL.
If enabled, the agent may prompt you daily and then use replies to log reading progress.
The skill can create a recurring scheduled check-in. It is opt-in and includes instructions to remove it, so this is disclosed persistence rather than hidden background behavior.
If the user asks for the reminder... Schedule a cron job... `openclaw cron add --name "SansFiction reading check-in" ... --deliver --channel last`
Enable the reminder only if you want recurring prompts, specify your preferred time/timezone, and use the documented disable command if you no longer want it.