Back to skill

Security audit

postfast

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this social-media management skill suspicious because it reports undeclared analytics features and incomplete scanner coverage, while several other assessed areas are rated okay.

The verdict is suspicious, not malicious, and overall confidence is medium. Before installing, consider whether you want the skill to access follower-history and post-analytics features beyond its stated purpose, and whether you are comfortable with its upload and publishing workflows. Scanner coverage is incomplete: 40 of 43 components were fully inspected, three were partial, and some instruction checks were degraded.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (120)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
93% confidence
Finding

The description does not declare follower analytics or historical audience metrics. Fetching follower history is a distinct capability, rather than a supporting implementation detail of the listed publishing and account-management operations. The API host and authentication method are consistent with PostFast, and no unrelated data transfer is evident.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
92% confidence
Finding

The request performs follower analytics retrieval, which is not included in the declared publishing, account-listing, or comment-management capabilities. Although it uses the expected PostFast API and authentication mechanism, follower-history analysis is a distinct undeclared capability rather than a supporting implementation detail of the stated operations.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
90% confidence
Finding

The description lists publishing and management capabilities but does not declare analytics retrieval. Querying a dedicated analytics endpoint is a distinct reporting capability, rather than an implementation detail of scheduling posts or checking their status. The API host and authentication are otherwise consistent with the declared service.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
89% confidence
Finding

The request uses the expected PostFast service and authentication, but retrieves analytics rather than scheduled-post records or another declared resource. Analytics reporting is a distinct capability not stated in the description; it is not merely an implementation detail of scheduling or managing posts.

Content

No source excerpt is available for this finding.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · SKILL.md (reported line 155)May include surrounding context.

md
Returns `200` (not 201) with `{ "deletedIds": [...], "notFoundIds": [...] }`. `notFoundIds` holds ids that don't exist or belong to another workspace; a repeated id is reported once, and repeating the same call is harmless. `ids` takes 1-100 UUIDs: `[]` returns `400` "At least one id is required", more than 100 returns `400` "ids cannot exceed 100 entries", a non-UUID returns `400` "Each id must be a valid UUID". One call counts as one request against this endpoint's own limit, separate from `DELETE /social-posts/:id`.

**Delete everything that matches a filter** (for example every scheduled post on one account, after the user confirmed): don't advance `page` while deleting, because later posts shift onto pages you have already read and get skipped. Instead repeat: read `GET /social-posts?socialMediaIds=ACCOUNT_ID&statuses=SCHEDULED&page=0&limit=50`, send the returned ids to `POST /social-posts/bulk-delete`, and stop when `data` comes back empty. Stop on any error too; after a `429`,

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 635)May include surrounding context.

md
he wording. Every field, error code and edge case: [references/social-inbox.md](references/social-inbox.md).

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 866)May include surrounding context.

md
- [references/social-inbox.md](references/social-inbox.md): The Social Inbox routes, objects and error codes. Generated from the docs

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · SKILL.md (reported line 672)May include surrounding context.

md
- The URL must be `https` on port 443 with a public hostname (no IP address, no `localhost`). A workspace can have 5 active endpoints.
- Posts imported from a platform when an account connects send no events, and neither does a first comment that fails.

**Routes:** `POST /webhook-endpoints` (create; body `url`, `events`, optional `description`), `GET /webhook-endpoints` (list), `GET /webhook-endpoints/{id}`, `PATCH /webhook-endpoints/{id}` (any of `url`, `events`, `description`, `active`), `DELETE /webhook-endpoints/{id}` (also deletes its delivery log), `GET /webhook-endpoints/{id}/deliveries` (the last 30 days: each attempt's status code, latency and error, and the payload as sent), `POST /webhook-endpoints/{id}/test`. Errors carry a `webhook.*` code in `message`, for example `webhook.urlNotHttps`, `webhook.urlHostNotAllowed`, `webhook.limitReached` and `webhook.notFound`.

Payload examples for each event, the retry schedule, every error code and the signature check in Node.js and P

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · references/webhooks.md (reported line 320)May include surrounding context.

md
ebhook-endpoints` | Lists the workspace's webhook endpoints, oldest first, disabled ones included. | 6,000 requests per day, 600 per hour |
| `GET /webhook-endpoints/{id}` | Fetches one webhook endpoint. Returns `404` once it has been deleted. | 6,000 requests per day, 600 per hour |
| `PATCH /webhook-endpoints/{id}` | Changes an endpoint's URL, events or description, or turns it off and on. `active: true` re-enables a disabled endpoint and resets its failure state. | 180 requests per hour |
| `DELETE /webhook-endpoints/{id}` | Deletes a webhook endpoint and its delivery log. Deliveries still waiting for a retry are dropped. | 180 requests per hour |
| `GET /webhook-endpoints/{id}/deliveries` | Lists an endpoint's deliveries from the last 30 days, newest first: the event as sent, and each attempt's status code, latency and error. | 6,000 requests per day, 600 per hour |
| `POST /webhook-endpoints/{id}/test` | Sends one signed delivery to the endpoint and returns the result: `webhook.te

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · SKILL.md (reported line 698)May include surrounding context.

md
| `POST /social-media/connect-link` | 180/hour |
| `GET /social-posts` | 216/minute, 540/5 minutes, 2,160/hour, 21,600/day |
| `POST /social-posts` | 180/minute, 1,260/day |
| `DELETE /social-posts/:id` | 600/hour, 6,000/day |
| `POST /social-posts/bulk-delete` | 600/hour, 6,000/day |
| `GET /social-posts/analytics` | 216/minute, 540/5 minutes, 3,780/hour, 37,800/day |
| `GET /social-inbox/conversations` | 360/minute, 900/5 minutes, 5,400/hour, 54,000/day |

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · SKILL.md (reported line 719)May include surrounding context.

md
| `GET /webhook-endpoints/:id/deliveries` | 600/hour, 6,000/day |
| `POST /webhook-endpoints/:id/test` | 10/minute, 60/hour, 600/day |

For example, `DELETE /social-posts/:id` sets 600/hour, 6,000/day, so it allows 72/minute, 180/5 minutes, 600/hour, 6,000/day.
<!-- END GENERATED rate-limits -->

**How the windows behave:** a window starts at the endpoint's first call, not on a clock boundary (the 24-hour window is not a midnight reset). Once a window is exceeded, that endpoint stays blocked for the full length of the window, and **calls made while blocked still count toward it**, so retrying in a loop keeps it blocked.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · SKILL.md (reported line 791)May include surrounding context.

md
rm**: `controls` apply to the whole batch and geo fields validate per platform, so put Facebook posts in one request and Instagram posts in another.
33. **Unknown fields and query parameters are ignored, not rejected**: A misspelled field, `status` inside a post object, or the singular `socialMediaId` on `GET /social-posts` returns no error; it simply has no effect. Use the exact names in this skill (`socialMediaIds`, plural, is the account filter).
34. **Deleting never touches the platform**: `DELETE /social-posts/:id` and `POST /social-posts/bulk-delete` remove posts from PostFast only. A published post stays live on the social network; the user removes it there.
35. **`totalCount` is `null` when nothing matches**, not `0`. Check `data.length` instead.
36. **Rate limits are per endpoint, and blocked calls count**: A `429` on one endpoint doesn't block the others. Wait `Retry-After` seconds; retrying sooner extends the block.
37. **TikTok sounds rotate and exclude each other**: `tikto

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · references/api-reference.md (reported line 24)May include surrounding context.

md
| `POST /social-media/connect-link` | Generates a secure connect link that lets someone connect their social accounts to your workspace without a PostFast account. | 180/hour | below |
| `GET /social-posts` | Query and paginate your social posts with filters. | 216/minute, 540/5 minutes, 2,160/hour, 21,600/day | below |
| `POST /social-posts` | Creates and schedules one or more social posts (up to 15 per request). | 180/minute, 1,260/day | below |
| `DELETE /social-posts/:id` | Deletes one social post, in any status. | 600/hour, 6,000/day | below |
| `POST /social-posts/bulk-delete` | Deletes up to 100 social posts in one request. | 600/hour, 6,000/day | below |
| `GET /social-posts/analytics` | Fetch published posts with their latest performance metrics. | 216/minute, 540/5 minutes, 3,780/hour, 37,800/day | below |
| `GET /social-inbox/conversations` | Lists comment conversations across your connected accounts, newest activity first. | 360/minute, 900/5 minutes, 5,400/hour, 54,000/day

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · references/api-reference.md (reported line 74)May include surrounding context.

md
- 360 requests per hour
- 2,400 requests per day

For example, `DELETE /social-posts/:id` sets 6,000 requests per day, 600 per hour, so it allows 6,000 requests per day, 600 per hour, 180 per 5 minutes, 72 per minute.

**Windows:** a window starts at the endpoint's first call, not on a clock boundary (the 24-hour window is not a reset at midnight UTC). Once a window is exceeded, that endpoint stays blocked for the full length of the window from that moment, and calls made while blocked still count toward it.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · SKILL.md (reported line 715)May include surrounding context.

md
| `GET /webhook-endpoints` | 600/hour, 6,000/day |
| `GET /webhook-endpoints/:id` | 600/hour, 6,000/day |
| `PATCH /webhook-endpoints/:id` | 180/hour |
| `DELETE /webhook-endpoints/:id` | 180/hour |
| `GET /webhook-endpoints/:id/deliveries` | 600/hour, 6,000/day |
| `POST /webhook-endpoints/:id/test` | 10/minute, 60/hour, 600/day |

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · references/api-reference.md (reported line 41)May include surrounding context.

md
| `GET /webhook-endpoints` | Lists the workspace's webhook endpoints, oldest first, disabled ones included. | 600/hour, 6,000/day | [webhooks.md](webhooks.md) |
| `GET /webhook-endpoints/:id` | Fetches one webhook endpoint. | 600/hour, 6,000/day | [webhooks.md](webhooks.md) |
| `PATCH /webhook-endpoints/:id` | Changes an endpoint's URL, events or description, or turns it off and on. | 180/hour | [webhooks.md](webhooks.md) |
| `DELETE /webhook-endpoints/:id` | Deletes a webhook endpoint and its delivery log. | 180/hour | [webhooks.md](webhooks.md) |
| `GET /webhook-endpoints/:id/deliveries` | Lists an endpoint's deliveries from the last 30 days, newest first: the event as sent, and each attempt's status code, latency and error. | 600/hour, 6,000/day | [webhooks.md](webhooks.md) |
| `POST /webhook-endpoints/:id/test` | Sends one signed delivery to the endpoint and returns the result: `webhook.test`, or a sample of any event with `sampleEvent`. | 10/minute, 60/hour, 600/day | [webhooks.md

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 865)May include surrounding context.

md
- [references/post-examples.md](references/post-examples.md): One complete `POST /social-posts` call per platform and post type. Generated from the docs

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · SKILL.md (reported line 930)May include surrounding context.

md
GET /social-posts?ids=POST_ID_1,POST_ID_2

# Delete posts (PostFast only: a published post stays live on the platform)
DELETE /social-posts/:id                            -> { deleted: true | false }
POST   /social-posts/bulk-delete  { ids: [1-100] }  -> { deletedIds, notFoundIds }

# Upload media (3 steps)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · references/api-reference.md (reported line 442)May include surrounding context.

md
**Request examples per platform and post type** (TikTok Video, Facebook Post, Facebook Reel, Facebook Story, Instagram Timeline, Instagram Story, Instagram Reel, TikTok Carousel, TikTok Draft, TikTok AI-Generated, Instagram AI-Labeled, Instagram Trial Reel, YouTube Shorts, X Retweet, Pinterest Pin, Threads Carousel, Instagram Carousel, X Post with First Comment, LinkedIn Document, Google Business Profile — Standard Update, Google Business Profile — Offer Post): [post-examples.md](post-examples.md).

### DELETE /social-posts/:id

Deletes one social post, in any status. It is removed from PostFast only: a published post stays live on the platform.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · SKILL.md (reported line 974)May include surrounding context.

md
GET    /webhook-endpoints
GET    /webhook-endpoints/:id
PATCH  /webhook-endpoints/:id        { url?, events?, description?, active? }
DELETE /webhook-endpoints/:id
GET    /webhook-endpoints/:id/deliveries
POST   /webhook-endpoints/:id/test   { sampleEvent? }

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · references/webhooks.md (reported line 465)May include surrounding context.

-d '{"active": true}'

text

### DELETE /webhook-endpoints/:id

Deletes a webhook endpoint and its delivery log. Deliveries still waiting for a retry are dropped.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · references/api-reference.md (reported line 467)May include surrounding context.

md
**Body:** `ids` (required, 1-100 post UUIDs, each a UUID v4). An empty array returns `400` (`At least one id is required`), more than 100 returns `400` (`ids cannot exceed 100 entries`), and a non-UUID returns `400` (`Each id must be a valid UUID`). **Response 200** (not 201): `{ "deletedIds": [...], "notFoundIds": [...] }`. `notFoundIds` holds the ids that do not exist or belong to another workspace (both get the same answer); a repeated id is reported once, and repeating a call with the same ids is harmless (they all come back in `notFoundIds`).

**What it deletes:** the same as `DELETE /social-posts/:id`: posts in any status, removed from PostFast only, so published posts stay live on the platform. One call counts as one request against this endpoint's own rate limit, separate from `DELETE /social-posts/:id`.

**cURL:**

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · references/api-reference.md (reported line 467)May include surrounding context.

md
**Body:** `ids` (required, 1-100 post UUIDs, each a UUID v4). An empty array returns `400` (`At least one id is required`), more than 100 returns `400` (`ids cannot exceed 100 entries`), and a non-UUID returns `400` (`Each id must be a valid UUID`). **Response 200** (not 201): `{ "deletedIds": [...], "notFoundIds": [...] }`. `notFoundIds` holds the ids that do not exist or belong to another workspace (both get the same answer); a repeated id is reported once, and repeating a call with the same ids is harmless (they all come back in `notFoundIds`).

**What it deletes:** the same as `DELETE /social-posts/:id`: posts in any status, removed from PostFast only, so published posts stay live on the platform. One call counts as one request against this endpoint's own rate limit, separate from `DELETE /social-posts/:id`.

**cURL:**

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

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).

Content

Scanner excerpt · references/api-reference.md (reported line 569)May include surrounding context.

md
## Webhooks

In [webhooks.md](webhooks.md): `POST /webhook-endpoints`, `GET /webhook-endpoints`, `GET /webhook-endpoints/:id`, `PATCH /webhook-endpoints/:id`, `DELETE /webhook-endpoints/:id`, `GET /webhook-endpoints/:id/deliveries`, `POST /webhook-endpoints/:id/test`.

## Error Responses

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 32)May include surrounding context.

1. List connected accounts

bash
curl -s -H "pf-api-key: $POSTFAST_API_KEY" https://api.postfa.st/social-media/my-social-accounts

Returns array of { id, platform, platformUsername, displayName, connectionStatus, disabledReason, inboxCapable, followerCount?, followerCountUpdatedAt? }. Save the id (the socialMediaId required for every post).