Security audit
Security checks for vulnerabilities and agentic risk
This is a low-risk WhatsApp formatting guide with some broad trigger and boilerplate wording that could cause unwanted formatting, but no evidence of hidden code, exfiltration, persistence, or destructive behavior.
Install only if you want an agent to format WhatsApp-targeted messages using WhatsApp syntax. Watch for accidental activation because the trigger words are broad, and prefer explicit use for WhatsApp messages rather than letting it affect unrelated writing.
The trigger keywords are extremely broad terms such as 'ensure', 'messages', and 'skill', which are likely to match normal conversation and activate the skill unintentionally. In an agent environment, overbroad activation can cause prompt-routing mistakes, unwanted output transformations, and interference with unrelated user tasks.
The repeated trigger list remains underspecified and overly generic, increasing the chance of accidental invocation across unrelated requests. Because this skill changes message formatting behavior, unintended activation could silently alter outputs or override user-preferred formatting in non-WhatsApp contexts.
The skill description imposes platform- and locale-specific behavior for all messages sent to WhatsApp without clearly requiring user opt-in. While not a direct exploit primitive, forced behavior can create integrity and usability issues by transforming content unexpectedly or conflicting with user language preferences and formatting needs.
Lines L053-L054 describe the skill as "MD+execute" and say some functions require exec command-line capability. However, the rest of the file is purely instructional content about WhatsApp text formatting, with no commands, scripts, or operational steps that actually invoke execution. This is an intent-level contradiction in the documentation about how the skill works.
L050 states the skill is based on Markdown instructions and needs no additional API key except where an external API is explicitly marked, but the skill contains no such external API integration. Later sections discuss network errors and possible external API limits, which suggests capabilities not actually present in the skill content. This creates contradictory operational expectations in the documentation.
No suspicious patterns detected.
