Back to skill

Security audit

Find Pr Agency

Security checks for vulnerabilities and agentic risk

Overview

The skill’s agency-search purpose is legitimate, but its documented fallback auth flow tells agents to execute a local .env file as shell code and can also send sensitive business intent to a third-party API without explicit minimization.

Review before installing. Prefer the ServiceGraph MCP/OAuth path, or require SERVICEGRAPH_API_KEY to already be present in the process environment. Do not let the agent source .env.local. For sensitive fundraising, IPO, crisis, legal, regulatory, healthcare, or unreleased launch details, ask the agent to build filters locally or confirm exactly what minimized text will be sent to ServiceGraph.

Vulnerability Patterns
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:81
Finding

Arbitrary Command Execution Through Dot-Sourcing of .env.local

Content
View full analysis
Remediation
View remediation

other

Warning
Location
SKILL.md:207
Finding

Potential Disclosure of Sensitive Business Intent to a Third-Party Translation Endpoint

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • 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
Findings (4)

Credential Access

High
Category
Privilege Escalation
Confidence
93% confidence
Finding

The skill instructs the agent to source .env.local during execution, which is a credential-access pattern that can expose secrets if tool outputs, command traces, errors, or follow-on prompts leak environment contents into model-visible context. In an adversarial or poorly isolated harness, encouraging access to local secret files expands the attack surface and can normalize secret-handling behavior that other malicious skills could exploit.

Content

Scanner excerpt · SKILL.md (reported line 85)May include surrounding context.

.env.local:

bash
( set -a; [ -f .env.local ] && . ./.env.local; set +a;
  curl -sS -H "Authorization: Bearer $SERVICEGRAPH_API_KEY" \
       'https://api.servicegraph.co/v1/datasets/pro_services/fields' )

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The description says the skill should trigger not only on explicit agency-finding requests but also 'even when described indirectly' with examples like 'we need press' and 'manage our brand reputation.' These phrases are broad business-language expressions that can overlap with requests for strategy, writing, or general advice, increasing the risk of invoking this skill outside its intended scope.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The documentation says to 'never read .env* files' and to keep the token out of LLM context, but the very next example instructs using a shell wrapper that sources .env.local. That is an active contradiction in the skill's own guidance about how credentials should be accessed and creates ambiguity about whether the agent should touch local secret files.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 87)May include surrounding context.

bash
   ( set -a; [ -f .env.local ] && . ./.env.local; set +a;
     curl -sS -H "Authorization: Bearer $SERVICEGRAPH_API_KEY" \
          'https://api.servicegraph.co/v1/datasets/pro_services/fields' )
  1. On 401 unauthorized, prompt the user: