Security audit

Work Productivity Pollyreach Gives Workflow Helper

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation-only workflow helper with overly broad activation wording but no hidden code, credential handling, persistence, or destructive behavior.

Install only if you want a general PollyReach-style workflow helper. Review or narrow its trigger keywords before enabling implicit invocation, because common words like phone, number, things, and done could cause unintended activation in unrelated conversations.

A.I.G

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

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (11)

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The trigger sentences are extremely broad and partially malformed, and they can match common user requests unrelated to this skill. In an agent-routing system, this increases the chance of unintended skill activation, causing the wrong workflow to run, misleading outputs, or inappropriate handling of user data and tasks.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The keyword list includes highly generic terms such as 'every', 'phone', 'number', 'ability', 'things', and 'done', which are likely to appear in many unrelated conversations. This creates a broad activation surface that can misroute requests to this skill, reducing system reliability and potentially exposing users to unintended processing paths.

Vague Triggers

High

Confidence: 96% confidence
Finding: The trigger keywords and example phrases are extremely broad, including generic terms like 'phone', 'number', 'ability', 'things', 'done', and 'bug fix'. This can cause the skill to activate for unrelated everyday requests, leading to unintended routing, prompt/context hijacking of normal tasks, and reduced trust in the agent's skill-selection boundary.

Vague Triggers

High

Confidence: 96% confidence
Finding: The skill description is written so broadly that it can match many ordinary productivity requests unrelated to this specific workflow. In an agent routing system, this increases unintended activation, which can hijack user requests, bypass more appropriate skills, and expand the skill's effective authority beyond its intended scope.

Vague Triggers

High

Confidence: 99% confidence
Finding: The trigger keyword list includes extremely generic words such as 'every', 'phone', 'number', 'ability', 'things', and 'done', which are likely to appear in a wide range of benign conversations. This makes accidental invocation highly probable and can cause prompt-routing collisions or skill shadowing across unrelated tasks.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The example trigger phrases begin with generic formulations like 'Help me' and 'I need a practical workflow' and provide little specificity about when this skill should be chosen. These examples can train or bias routing toward activating the skill for common requests, reinforcing over-broad matching behavior.

Vague Triggers

High

Confidence: 95% confidence
Finding: The trigger keywords include very common terms such as "every", "phone", "number", "ability", and "things", which can cause the skill to activate in unrelated conversations. In an agent environment, overbroad activation can route user requests to the wrong workflow, producing irrelevant actions, unintended data handling, or bypass of more appropriate safeguards.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The activation description says to use the skill when users mention broad topics like work-productivity or need "practical workflow, artifact, checklist, analysis, or implementation support," which is too vague to establish safe and predictable routing boundaries. This ambiguity increases the chance of accidental invocation and misapplication of the skill to requests outside its intended scope.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The skill enables implicit invocation without any visible activation constraints, examples, or narrowing conditions. This can cause the agent to auto-select the skill in loosely related conversations, increasing the chance of unintended execution, overreach into user workflows, and misuse in contexts the skill was not designed for.

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The default prompt includes broad, everyday trigger language such as 'help me' and a wide set of generic work-productivity terms, which makes accidental matching more likely. Combined with the productivity-oriented scope, this broad phrasing can pull the skill into many unrelated conversations and amplify unintended behavior or policy bypass through over-selection.

Vague Triggers

High

Confidence: 97% confidence
Finding: The trigger phrases are excessively broad and include common words and fragments such as "gives," "every," "phone," "number," and "ability," which can cause the skill to activate for unrelated everyday requests. This creates prompt-routing ambiguity and increases the chance that the wrong skill is invoked, potentially producing irrelevant actions, misleading outputs, or unsafe workflow handling in contexts the skill was not designed for.

Static analysis

No suspicious patterns detected.