Back to skill

Security audit

Compact Guard

Security checks for vulnerabilities and agentic risk

Overview

The skill is coherent and documentation-only, but it persists broad conversation and tool-output details to memory without a clear review, redaction, or opt-out gate.

Review before installing if your sessions may contain credentials, private endpoints, personal information, customer data, or sensitive tool output. Use it only when you are comfortable with extracted details being written to memory files, and prefer adding an explicit review/redaction step before any memory append.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:34
Finding

Overbroad Persistence of Potentially Sensitive Conversation Data

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Note
Location
README.md:8
Finding

Installation Instructions Use Unpinned Mutable Sources

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The documented trigger phrases are broad and lack any activation constraints, scoping, or negative examples, which increases the chance of accidental or unintended invocation. In a skill that performs scanning, extraction, memory-file writes, and prepares for compaction, ambiguous triggers can cause the agent to run state-changing pre-compaction steps in the wrong context.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill explicitly says it should proactively suggest or trigger when context is near full, and it lists broad natural-language trigger phrases. In a system where skills may be auto-invoked from conversational cues, this ambiguity can cause the skill to run unexpectedly, leading to unsolicited scanning of conversation/tool output and preparation for persistence actions.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill instructs the agent to extract facts, preferences, errors, and progress from conversation and tool outputs, then append them to persistent memory files, but it does not present this as an explicit user-facing data persistence warning at the point of use. This creates a privacy and data-retention risk because users may not realize potentially sensitive content from transient context will be stored durably.

Content

No source excerpt is available for this finding.