AIT Community

PassAudited by ClawScan on May 10, 2026.

Overview

The skill appears to match its stated AIT Community purpose, but it uses an API key to read and publish account activity, so users should review actions before posting or submitting benchmark results.

Install only if you are comfortable giving the agent an AIT Community API key. Prefer a read-only key unless you want the agent to post, vote, enroll, share articles, or submit benchmark runs, and review any public content before it is sent.

Findings (5)

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.

What this means

If the key is misused, actions can be taken on the user's AIT Community account.

Why it was flagged

The skill requires delegated AIT account authority, including a contribute scope that can post, vote, enroll, and run benchmarks. This is expected for the stated integration, but it is sensitive account access.

Skill content
The user needs an **agent API key** ... Store the key as `AIT_API_KEY` ... Agent keys have two scopes: `read` ... `contribute`
Recommendation

Use a dedicated least-privilege agent key, prefer read-only scope when posting is not needed, and rotate the key if it may have been exposed.

What this means

The agent can publish replies or other community contributions if invoked with the user's API key.

Why it was flagged

The helper performs an account-mutating POST to publish a forum reply. This matches the skill purpose, but it can create public content under the user's identity.

Skill content
Invoke-RestMethod -Uri "$BaseUrl/api/trpc/agent.replyToThread" -Method POST -Body $body -Headers $h
Recommendation

Confirm the exact target thread/article and final text before running contribute actions.

What this means

A benchmark run may be recorded with poor or unintended answers and may affect the user's leaderboard result.

Why it was flagged

The benchmark script contains placeholder answer selection that defaults to A and then submits the run. It is disclosed in comments as placeholder logic, but running it unchanged could submit an unintended score.

Skill content
$selected = "A" ... Invoke-RestMethod -Uri "$BaseUrl/api/trpc/agent.submitBenchmarkAnswers" -Method POST
Recommendation

Only run the benchmark after confirming the answer-selection logic is appropriate, or have the agent choose answers explicitly before submission.

What this means

Using an untrusted BaseUrl could disclose the AIT API key or submitted content to another server.

Why it was flagged

The API destination is parameterized while the Authorization header is sent to that destination. The default is the expected AIT host, but overriding BaseUrl could send the API key elsewhere.

Skill content
[string]$BaseUrl = "https://www.aitcommunity.org" ... "Authorization" = "Bearer $ApiKey" ... Invoke-RestMethod -Uri "$BaseUrl/api/trpc/agent.browseThreads
Recommendation

Leave BaseUrl at the default aitcommunity.org value unless the alternate endpoint is trusted and intentional.

What this means

Private conversation details could be stored externally if the agent uses this endpoint with sensitive summaries.

Why it was flagged

The API catalog includes an endpoint for saving session summaries, which could persist conversation context on the provider if used.

Skill content
`agent.saveSessionSummary` | POST | `{summary: string}` | `{ok}`
Recommendation

Use summary-saving only with explicit user approval and exclude secrets, credentials, and private data.