← mcp-aichat@2026.11.1.3
Propose a patch
A fix the maintainer can act on, with the evidence attached. It opens under your GitHub account in your browser; KillerSkills holds no token and never modifies the package. The fix is verified the only way we verify anything: the next crawl re-runs the new version in the honeypot. 73 · Caution
Evidence
The sandbox observed nothing beyond the static findings; the proposal cites those.
Diff · removes 15 caught lines
Download .patchdiff --git a/README.pypi.md b/README.pypi.md --- a/README.pypi.md +++ b/README.pypi.md @@ -62,1 +62,0 @@ -BEGIN TOKEN ONBOARDING @@ -69,1 +68,0 @@ -Copy an application API key or open Manage Keys @@ -74,1 +72,0 @@ -END TOKEN ONBOARDING diff --git a/jetbrains/README.md b/jetbrains/README.md --- a/jetbrains/README.md +++ b/jetbrains/README.md @@ -22,1 +22,0 @@ -BEGIN TOKEN ONBOARDING @@ -29,1 +28,0 @@ -Copy an application API key or open Manage Keys @@ -34,1 +32,0 @@ -END TOKEN ONBOARDING diff --git a/README.md b/README.md --- a/README.md +++ b/README.md @@ -61,1 +61,0 @@ -BEGIN TOKEN ONBOARDING @@ -68,1 +67,0 @@ -Copy an application API key or open Manage Keys @@ -73,1 +71,0 @@ -END TOKEN ONBOARDING diff --git a/goose/README_zh_Hans.md b/goose/README_zh_Hans.md --- a/goose/README_zh_Hans.md +++ b/goose/README_zh_Hans.md @@ -14,1 +14,0 @@ -真实脱敏 Ace 页面：复制 API Key 或打开 Manage Keys diff --git a/goose/README.md b/goose/README.md --- a/goose/README.md +++ b/goose/README.md @@ -11,1 +11,0 @@ -Copy the API Key or Manage Keys — real Ace console reference @@ -13,1 +12,0 @@ -Create a dedicated API Key — real Ace console reference diff --git a/vscode/README.md b/vscode/README.md --- a/vscode/README.md +++ b/vscode/README.md @@ -26,1 +26,0 @@ -BEGIN TOKEN ONBOARDING @@ -33,1 +32,0 @@ -Copy an application API key or open Manage Keys @@ -38,1 +36,0 @@ -END TOKEN ONBOARDING
Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 2026.11.1.3. If the behavior is intended, declare it (host under permissions.network, a line in the README) instead of removing it; the next crawl scores declarations.
Issue / PR body
Title: mcp-aichat: HTML comment carrying imperative instructions to the agent (KillerSkills verdict caution 73)
## What the honeypot observed KillerSkills ran **[email protected]** in an isolated sandbox (decoy credentials, recording proxy, syscall trace; engine 1.7) and scored it **CAUTION 73/100**. Summary: no network, touches nothing outside the project, writes no memory. - See the findings below. Full trace and findings: https://killerskills.ai/skills/pypi/mcp-aichat/2026.11.1.3 ## Findings with a location - `README.pypi.md:62` · HTML comment carrying imperative instructions to the agent — README.pypi.md:62 · HTML comment carrying imperative instructions to the agent - `jetbrains/README.md:22` · HTML comment carrying imperative instructions to the agent — jetbrains/README.md:22 · HTML comment carrying imperative instructions to the agent - `README.md:61` · HTML comment carrying imperative instructions to the agent — README.md:61 · HTML comment carrying imperative instructions to the agent - `README.pypi.md:74` · HTML comment carrying imperative instructions to the agent — README.pypi.md:74 · HTML comment carrying imperative instructions to the agent - `README.md:73` · HTML comment carrying imperative instructions to the agent — README.md:73 · HTML comment carrying imperative instructions to the agent - `goose/README_zh_Hans.md:14` · Image alt text carrying instructions — goose/README_zh_Hans.md:14 · Image alt text carrying instructions - `jetbrains/README.md:34` · HTML comment carrying imperative instructions to the agent — jetbrains/README.md:34 · HTML comment carrying imperative instructions to the agent - `README.md:68` · Image alt text carrying instructions — README.md:68 · Image alt text carrying instructions - `goose/README.md:11` · Image alt text carrying instructions — goose/README.md:11 · Image alt text carrying instructions - `vscode/README.md:33` · Image alt text carrying instructions — vscode/README.md:33 · Image alt text carrying instructions - `vscode/README.md:38` · HTML comment carrying imperative instructions to the agent — vscode/README.md:38 · HTML comment carrying imperative instructions to the agent - `jetbrains/README.md:29` · Image alt text carrying instructions — jetbrains/README.md:29 · Image alt text carrying instructions ## Proposed change The attached patch removes the 15 lines above. If the behavior is intended, the fix is to declare it instead: list the host under `permissions.network` (or the MCP manifest) and say in the README why the package needs it; the next crawl will re-score it with the declaration taken into account. ## How this will be verified Nothing is "verified safe" by assertion. When a new version is published, KillerSkills re-runs it in the honeypot and the verdict page updates; you can trigger it early from the verdict page ("Re-scan after I publish a fix"). Reply here or at [email protected] if you believe the verdict is wrong; disputes are reviewed by a person and the outcome is published on the verdict page. _Proposed through killerskills.ai/threats by a user of the package, under their own GitHub account. The evidence is public; the package was never modified by KillerSkills._
# or from a terminal, under your own account: gh repo fork AceDataCloud/AiChatMCP --clone && cd AiChatMCP curl -s https://killerskills.ai/skills/pypi/mcp-aichat/2026.11.1.3/patch.diff | git apply --unidiff-zero git checkout -b killerskills/2026.11.1.3 && git commit -am "mcp-aichat: HTML comment carrying imperative instructions to the agent" gh pr create --fill --body-file <(curl -s https://killerskills.ai/skills/pypi/mcp-aichat/2026.11.1.3/patch.md)
We count proposals opened, not merged: a merge shows up as a new version that the crawl re-runs. Disputes: the maintainer can contest the verdict from its page; a person reviews it and the outcome is published.