Security audit

HF Daily Researcher Min

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a disclosed research-report workflow, but it asks to use persistent user memory and can upload generated reports to Feishu without a clear consent boundary.

Install only if you are comfortable with a research workflow that may inspect persistent user memory for topics, spawn multiple agents, save intermediate files and reports, and potentially publish reports to Feishu. Prefer explicit topic entry, local-only report saving, and per-run confirmation before any cloud upload or cron scheduling.

A.I.G

Vulnerability Patterns

Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review

Findings (2)

T05 · Unauthorized Access and Privilege Escalation

Warning

Location: SKILL.md:496
Finding: Overbroad Access to Persistent User Memory for Research-Topic Discovery
Content: View full analysis
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Warning

Location: SKILL.md:509
Finding: Optional Feishu Report Upload Lacks an Explicit Per-Run Consent Boundary
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 88% confidence
Finding: The skill auto-selects deep research mode using broad natural-language triggers such as 'survey', '半年', or '过去一年', which can be matched ambiguously and cause a much broader workflow than the user intended. In this skill, deep mode expands time range, spawns many sub-agents, and increases data collection and persistence, so a misclassification can materially increase exposure, cost, and unintended processing.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The skill states that reports are saved locally but elsewhere indicates they may also be uploaded to Feishu, creating an undeclared external data flow. Research reports can contain user interests, derived summaries, and possibly sensitive working context, so sending them to a third-party platform without explicit, just-in-time consent creates a real confidentiality and privacy risk.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill describes saving reports locally and uploading them to Feishu without a clear warning about persistence duration, storage location, or external sharing. Because the workflow aggregates research interests, paper selections, and generated analyses, this creates a real data-handling and consent issue, especially when used in enterprise or confidential research contexts.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Low

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The tool dependency table documents browser as the fallback for web_fetch, implying a browser-based final fallback path. But the concrete run_search example only shows fallback from kimi_search to web_fetch and contains no browser fallback, so the documentation overstates what the shown implementation does.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.