Security audit
Security checks for vulnerabilities and agentic risk
This skill coherently documents a cloud form service for collecting human input, with expected external API use, optional webhooks, polling, and deletion controls disclosed.
Before installing, confirm that you are comfortable sending form content, responses, uploaded files, and recipient metadata to letsclarify.ai and to any webhook URL you configure. Store the API key and delete tokens securely, approve any recurring cron polling intentionally, and confirm before deleting a form because that removes collected data.
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
`POST /api/v1/register` with `{"name": "...", "email": "..."}` → `{"api_key": "lc_...", "key_prefix": "lc_xxxxx"}`. Store securely, shown only once.
`DELETE /api/v1/register` with auth header → `{"deleted": true}`.
### Create Form
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
### Delete Form
`DELETE /api/v1/forms/{form_token}` with `X-Delete-Token: {delete_token}` → `{"deleted": true}`. Permanently removes form, submissions, and files.
### Webhooks
This markdown file documents the delete_form tool as permanently deleting a form and all its data, but it does not include a clear caution or user-facing warning about the irreversible nature of that action. Under the markdown-file criteria, destructive behaviors affecting user data should be accompanied by explicit warnings.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
| `get_results` | Yes | Fetch submissions with pagination and filters |
| `delete_form` | Yes | Permanently delete a form and all its data |
## REST API — Full Curl Examples
### Register an API Key
This section gives a concrete DELETE request for removing a form, and elsewhere the document states expired data is permanently deleted, but the deletion example itself lacks a prominent warning about irreversible loss of submissions and form state. Markdown guidance for destructive operations should clearly disclose the impact on user data.
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
1. **Create form** → save `form_token` and `api_key`
2. **Send URLs** to humans via Telegram, email, etc.
3. **Set up polling** — create a cron job that checks the summary every 5–15 minutes
4. **On poll**: check `submitted_total` vs `known_total`
- If all responded → fetch full results, act on them, delete cron job
- If not all → do nothing, wait for next poll
The markdown describes distributing unique recipient URLs via third-party channels and sending submission payloads to a configured webhook, including response_json, but does not warn that human-provided data will be transmitted through external services. Because this behavior can affect user privacy and data handling, the skill description should explicitly disclose it.
The section provides a direct API example for deleting an API key, but does not warn readers that this will revoke future authenticated access for agents using that key. Because this markdown describes a potentially disruptive action affecting system access, a brief warning would improve user disclosure.
No suspicious patterns detected.