Security audit
Security checks for vulnerabilities and agentic risk
This marketing-advice skill is coherent and does not include executable code, but it does use local preference and memory files for personalization.
Install this if you want reusable CMO-style marketing guidance that remembers your stated business facts and preferences. Before using it, review what is stored in ~/Clawic/data/cmo/ and ~/Clawic/profile.yaml, because the skill may load that context automatically and may save new preferences or business facts you reveal during work.
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
last: YYYY-MM-DD
## Company
<!-- What they sell, to whom, stage, team size -->
## Funnel Numbers (as stated by the user)
<!-- ACV, win rate, lead→opp rate, sales cycle, current lead volume, CAC, payback -->
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
## Original Research
The most reliable link and press magnet in B2B: run a survey or mine your own product data for a statistic nobody else can publish. Requirements: a defensible method (sample size and question wording published), a number that is genuinely new, and a single citable chart. Repeat annually — the second edition earns more than the first because it becomes a series people wait for. Never publish a proprietary-data statistic without checking customer contracts and privacy commitments (`compliance.md`).
## Visibility in AI Answer Engines
The skill explicitly instructs the agent to read ~/Clawic/data/cmo/config.yaml, ~/Clawic/profile.yaml, and ~/Clawic/data/cmo/memory.md automatically on first use, without any user-facing notice or consent step. Even though these files are used for personalization, they may contain sensitive personal, business, or historical context data, so silent access increases privacy risk and can expose more data than is necessary for the immediate task.
No suspicious patterns detected.