Security audit
Security checks for vulnerabilities and agentic risk
This is a plain Markdown skill for drafting and reviewing product requirements, with no bundled scripts, persistence, credential handling, or hidden automation.
Installers should treat this as a drafting aid, not an authority for final product, legal, payment, privacy, reward, or compliance decisions. It may create local PRD or HTML demo files when asked and may use private templates if you provide them, so review generated requirements before sharing or implementing them.
Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.
| Function name | |
| User story | |
| Trigger condition | |
| Display rule | |
| Operation rule | |
| Data rule | |
| Permission rule | |
Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.
- Status tabs or state switcher for not started, ongoing, reserved, claimed, and ended
- Admin fields for campaign time, reward name, inventory, rules, and publish status
## HTML Demo Output Rules
If file creation is available:
The description says to use the skill when users ask to "analyze, break down, draft, review, standardize, prototype, or visualize product requirements," which is a very wide set of generic verbs without explicit trigger boundaries or exclusions. In a manifest-style file, this can cause unintended activation for ordinary drafting or review requests that may not actually be about this specific skill's intended scope.
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.
## Quality Guardrails
- Do not infer hidden rules as facts.
- Do not assume payment, reward inventory, personal data, or compliance rules without confirmation.
- Do not describe invisible image areas as if they are present.
- Keep screenshot analysis useful for PRD drafting, not just visual description.
- For competitor references, avoid copying brand assets, logos, or proprietary copy into final deliverables.
Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.
Software, and to permit persons to whom the Software is furnished to do so.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
The documentation states that no scripts, network calls, secrets, or executable installers are included, which reads as a strong safety/intent claim about the skill package. However, the same README later instructs users to run executable installer and management commands such as $skill-installer install, openclaw skills install, and openclaw skills check, so the documentation overstates the absence of executable tooling around the skill.
This markdown file applies to SQP-1, and its activation guidance is broad: it says to use the reference whenever a requirement belongs to several large domains and to 'choose the nearest scenario' without defining clear boundaries between when this reference should or should not be used. The absence of explicit exclusions or negative examples could cause unintended or inconsistent invocation for adjacent product requirements.
No suspicious patterns detected.