Fleet PR Agent

• Outcome: Review
• Latest audit: Sep 16, 2026 · 3:22 AM UTC
• Version: 1.1.0

Overview

The skill mostly does the advertised GitHub PR triage, but an included shell script has a real command-execution weakness if run with attacker-controlled environment variables.

Review before installing or running in automation. Prefer the documented Python script, avoid running scripts with untrusted environment variables, and require integer validation for FLEET_PR_STALE_DAYS, FLEET_PR_CI_WEIGHT, and FLEET_PR_MAX_PRS before using the Bash helper.

A.I.G

Vulnerability Patterns

• Insecure Skill Coding Practices: Finds exploitable flaws such as hardcoded secrets or command injection
• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally

Findings (1)

T09 · Insecure Skill Coding Practices

Error

• Location: scripts/triage.sh:94

Finding

Command Execution Through Unvalidated Bash Arithmetic Input

Content

View full analysis

Remediation

View remediation

&2 exit 2 fi } validate_uint "FLEET_PR_STALE_DAYS" "$STALE_DAYS" validate_uint "FLEET_PR_CI_WEIGHT" "$CI_WEIGHT" validate_uint "FLEET_PR_MAX_PRS" "$MAX_PRS" ``` After syntax validation, enforce sensible bounds to prevent denial-of-service or unintended behavior: ```bash if (( STALE_DAYS > 3650 )); then echo "Error: FLEET_PR_STALE_DAYS is outside the permitted range." >&2 exit 2 fi if (( MAX_PRS < 1 || MAX_PRS > 1000 )); then echo "Error: FLEET_PR_MAX_PRS must be between 1 and 1000." >&2 exit 2 fi ``` Additional hardening measures: 1. Validate configuration immediately after reading the environment and before invoking `date`, `gh`, or any arithmetic comparison. 2. Reject malformed values rather than attempting to normalize or evaluate them. 3. Apply the same numeric-validation policy to every current and future environment-controlled arithmetic value. 4. In CI and agent deployments, use an explicit environment allowlist and avoid forwarding untrusted environment variables to the Skill. 5. Add regression tests using malformed arithmetic strings and array-subscript payloads to verify that validation rejects them before PR processing begins. ]]>

SkillSpector

Vulnerability Patterns

• Behavioral AST: exec() Call, eval() Call, Dynamic Import
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (2)

Undeclared Tool Scope

Medium

• Category: MCP Least Privilege
• Confidence: 93% confidence

Finding

The skill advertises and operationalizes shell execution (gh, python3) and file output (--output report.md) but does not declare any explicit tool scope such as permissions or allowed-tools. This creates an authorization ambiguity where an agent runtime may permit broader shell or file access than is necessary, increasing the chance of unintended command execution, data access, or writes beyond the intended triage workflow.

Content

No source excerpt is available for this finding.

subprocess module call

Medium

• Category: Dangerous Code Execution
• Confidence: 70% confidence

Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/triage.py (reported line 22)May include surrounding context.

python

cmd = ["gh", "pr", "list", "--repo", repo, "--state", "open",
           "--limit", str(MAX_PRS), "--json",
           "number,title,url,createdAt,updatedAt,isDraft,reviewDecision,statusCheckRollup,labels"]
    result = subprocess.run(cmd, capture_output=True, text=True)
    if result.returncode != 0:
        return []
    try:

Static analysis

No suspicious patterns detected.