Security audit
Security checks for vulnerabilities and agentic risk
This skill is a research-manuscript writing assistant whose local references and validation scripts match that purpose, with no hidden network, credential, destructive, or privileged behavior found.
Install this if you want help drafting or auditing research manuscripts from your own materials. Treat unpublished manuscripts, data, target-paper notes, and journal models as sensitive inputs, and review generated prose for scientific judgment before using it.
This second mismatch finding highlights that the skill appears to validate local JSON models rather than simply provide manuscript guidance. Hidden operational behavior increases the chance that downstream systems grant access to sensitive files or trust outputs under false assumptions about what the skill does.
This second mismatch finding highlights that the skill appears to validate local JSON models rather than simply provide manuscript guidance. Hidden operational behavior increases the chance that downstream systems grant access to sensitive files or trust outputs under false assumptions about what the skill does.
Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.
## Output rules
- Put useful work before process explanation.
- Preserve good existing prose instead of forcing cosmetic rewrites.
The skill explicitly instructs the agent to read multiple local reference, asset, and script files, but it declares no tool scope or permissions. That creates an authority mismatch: a caller may invoke the skill expecting a self-contained prompt, while the skill actually depends on filesystem access, which can broaden data exposure and make review of reachable content incomplete.
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.
- facts, numbers, citations, technical terms, null findings, limitations, and author judgments that must be protected;
- the next useful output that can be produced safely now.
Do not ask the user to choose an internal mode. Do not require field, journal, section, or language preferences when a conservative useful result is possible.
If missing information would force an unsupported scientific choice:
The instruction says not to require language preferences when a conservative useful result is possible. This can cause the skill to proceed in a default language rather than offering the user a language choice, which matches the policy concern about forcing a specific language or locale without opt-in.
This markdown file includes the example request Help me write my paper, which is a broad natural-language phrase that overlaps with ordinary user speech. The document does not provide tighter trigger constraints or exclusion examples nearby, so the invocation scope may be interpreted too broadly.
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.
- starting with a universal claim too broad for the evidence;
- literature-by-literature summary without synthesis;
- gap inflation such as `no studies have examined` without verification;
- a study aim that does not answer the stated gap;
- adding hypotheses because the section appears incomplete;
- changing a user-supplied exploratory question into a confirmatory prediction.
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.
- starting with a universal claim too broad for the evidence;
- literature-by-literature summary without synthesis;
- gap inflation such as `no studies have examined` without verification;
- a study aim that does not answer the stated gap;
- adding hypotheses because the section appears incomplete;
- changing a user-supplied exploratory question into a confirmatory prediction.
Suspicious Unicode normalization or mixed-script content
No suspicious patterns detected.