Back to skill

Security audit

agent-responsiveness-delegation

Security checks for vulnerabilities and agentic risk

Overview

This skill is a clearly disclosed local setup and audit tool that changes supported agent instruction files only when explicitly applied.

Install only if you want this skill to persist delegation rules into your agent's instruction files. Run the provisioning script first without --apply to review the target files, keep the generated backups, and read the Chinese rule block before applying it globally.

Vulnerability Patterns
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (7)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill describes scripts that detect local agent environments and write to instruction/config files, but it does not declare any explicit tool scope or allowed-tools constraints. That mismatch increases the chance an agent will grant broader file and environment access than users expect, enabling unintended modification of local files or configuration through a loosely scoped skill.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The description lists trigger phrases such as “AI不回话”, “等太久”, “助手响应慢”, and “响应速度”, which are generic expressions that can appear in many ordinary conversations unrelated to this specific skill. The file does not provide negative examples or clear exclusion conditions to distinguish when the skill should activate versus when it should not.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The section says to use the skill when a user complains about waiting too long, interruptions being blocked, or wanting to chat frequently, but these conditions are high-level symptoms rather than precise invocation criteria. Without boundaries, the skill could be invoked in situations where the user is only expressing frustration, not asking for this configuration workflow.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The file is written as normative skill guidance entirely in Chinese and presents its instructions as the authoritative implementation reference. There is no indication that users may opt into another language or that the Chinese-only presentation is required for a region-specific purpose.

Content

No source excerpt is available for this finding.

Tainted flow: 'new' from pathlib.Path.read_text (line 242, file read) → pathlib.Path.write_text (file write)

Medium
Category
Data Flow
Confidence
65% confidence
Finding

Data from a source is assigned to a variable that is later passed to a sink, creating a variable-mediated taint flow.

Content

Scanner excerpt · scripts/provision_agent_responsiveness.py (reported line 257)May include surrounding context.

python
path.parent.mkdir(parents=True, exist_ok=True)
            if exists:
                rec["backup"] = str(backup(path, stamp))
            path.write_text(new, encoding="utf-8", newline="")
            applied = True
        actions.append(rec)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The rule block is entirely written in Chinese and presents itself as operational policy to be inserted into system prompts or custom instructions, with no indication that language should follow user preference. Inserting fixed-language governance text can degrade usability, cause misunderstanding of safety/operational rules by users or operators who do not read Chinese, and effectively impose a locale without opt-in.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

The manifest describes a local, zero-dependency skill for provisioning and auditing agent instruction files. The author-contact section adds external social/contact channels and a remote image URL, which are not justified by the operational purpose of configuring agent responsiveness and extend the skill content into self-promotion.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.