Security audit
Security checks for vulnerabilities and agentic risk
The skill is coherent with its stated purpose, but it should be reviewed because it autonomously posts financial signals and promotional content to Telegram on a schedule without a clear approval or dry-run boundary.
Install only if you intentionally want an autonomous agent to publish market signals, VIP upsells, and owner reports through your Telegram bot. Verify channel IDs, bot permissions, and compliance requirements first, and keep cron posting disabled or in a manual-review flow until you have checked the exact messages it will send.
SKILL.md:192Mandatory Promotional Output Injection Through Persistent Skill Instructions
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
[ ] tracker_state.json initialized with empty arrays
[ ] Public Telegram channel created and bot is admin
[ ] VIP Telegram channel created and bot is admin
[ ] PUBLIC_CHANNEL_ID set in .env
[ ] VIP_CHANNEL_ID set in .env
[ ] OWNER_CHAT_ID set in .env (for weekly reports)
[ ] Landing page URL defined (even a simple page works)
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
[ ] tracker_state.json initialized with empty arrays
[ ] Public Telegram channel created and bot is admin
[ ] VIP Telegram channel created and bot is admin
[ ] PUBLIC_CHANNEL_ID set in .env
[ ] VIP_CHANNEL_ID set in .env
[ ] OWNER_CHAT_ID set in .env (for weekly reports)
[ ] Landing page URL defined (even a simple page works)
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
[ ] tracker_state.json initialized with empty arrays
[ ] Public Telegram channel created and bot is admin
[ ] VIP Telegram channel created and bot is admin
[ ] PUBLIC_CHANNEL_ID set in .env
[ ] VIP_CHANNEL_ID set in .env
[ ] OWNER_CHAT_ID set in .env (for weekly reports)
[ ] Landing page URL defined (even a simple page works)
The README describes automated publishing of content and trading signals to public and VIP Telegram channels without a clear operator-facing warning about outbound posting. This creates risk of unintended disclosure, spam, reputational harm, regulatory exposure, or accidental publication of low-quality or sensitive content, especially because the skill is marketed as autonomous and scheduled. In this financial-signals context, automated public messaging is more dangerous because posts may be interpreted as investment guidance and can trigger real-world consequences immediately.
The README states that the skill will auto-create required files on first run, but it does not clearly warn users that the workspace will be modified. Silent file creation can surprise operators, overwrite expected layouts, or create artifacts that later influence agent behavior and downstream automation. In an autonomous-agent context, undisclosed write behavior is a meaningful safety concern even if it is framed as setup convenience.
The metadata declares outbound network access to Telegram and public APIs and requires a Telegram bot token, but it does not include a clear privacy or data-transmission warning describing what information may be sent externally. This is risky because an autonomous agent may transmit generated content, channel identifiers, market summaries, or operational metadata to third-party services without the user's fully informed consent.
The skill explicitly instructs the agent to create directories and files and to perform recurring writes across multiple workspace locations, but it does not present a clear user-facing warning or consent boundary before modifying local data. This is dangerous because autonomous file creation and continuous state logging can alter, overwrite, or clutter workspace data in ways the operator may not expect, especially when scheduled cron-like execution is implied.
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
</system_instructions>
<skill_reference>
Read and follow: /workspace/skills/agent-shark-mindset/SKILL.md
Execute: MODE 1 — DAILY ALPHA SCAN
</skill_reference>
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
</system_instructions>
<skill_reference>
Read and follow: /workspace/skills/agent-shark-mindset/SKILL.md
Execute: MODE 1 — DAILY ALPHA SCAN
</skill_reference>
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
</system_instructions>
<skill_reference>
Read and follow: /workspace/skills/agent-shark-mindset/SKILL.md
Execute: MODE 1 — DAILY ALPHA SCAN
</skill_reference>
The cron message explicitly authorizes autonomous publication to public/VIP Telegram channels and local writes without any user-facing disclosure, approval gate, or safety interlock. In this skill context, those side effects are central to operation and can cause unauthorized financial promotions, data leakage, or persistent state changes if the agent is misconfigured, compromised, or produces incorrect output.
The README states 'Shark tone enforced,' which imposes a fixed communication style on generated content. This is a natural-language policy concern because it removes user choice over output style without indicating any opt-in or override.
The skill config instructs the agent to send automated owner notifications over Telegram but does not warn about privacy, authenticity, or integrity risks of sending operational summaries through a third-party messaging channel. These messages may expose trading activity, revenue data, or workflow state, and could be misdelivered if the chat ID is wrong or if Telegram access is compromised.
No suspicious patterns detected.