Create Skills

• Outcome: Pass
• Latest audit: Sep 11, 2026 · 6:46 AM UTC
• Version: 0.1.2

Overview

This skill is a disclosed helper for creating or updating skill documentation, with no bundled executable code or hidden runtime behavior found.

Install this only if you want an agent helper that can create and revise skill definitions. Review proposed changes carefully when it updates triggers, coordinator files, agents, package metadata, or references, because those edits can change when future skills activate.

A.I.G

Vulnerability Patterns

• 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
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Rogue Agent: Self-Modification, Session Persistence
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (12)

Self-Modification

High

• Category: Rogue Agent
• Confidence: 95% confidence

Finding

The frontmatter description advertises the ability to 'update SKILL.md', which signals a capability to modify agent skills themselves. Because skill metadata controls invocation and behavior, this creates a persistence and privilege-escalation risk if misused to broaden triggers, alter allowed tools, or change coordination logic.

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.

md

---
name: document-skills
description: Write or update a skill (SKILL.md and supporting files) to match host structure and best practices. Use when user says create skill, write skill, update SKILL.md, /document-skills.
disable-model-invocation: true
argument-hint: "[skill-path] [source]"
context: fork

Self-Modification

High

• Category: Rogue Agent
• Confidence: 95% confidence

Finding

The frontmatter description advertises the ability to 'update SKILL.md', which signals a capability to modify agent skills themselves. Because skill metadata controls invocation and behavior, this creates a persistence and privilege-escalation risk if misused to broaden triggers, alter allowed tools, or change coordination logic.

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.

md

---
name: document-skills
description: Write or update a skill (SKILL.md and supporting files) to match host structure and best practices. Use when user says create skill, write skill, update SKILL.md, /document-skills.
disable-model-invocation: true
argument-hint: "[skill-path] [source]"
context: fork

Ae1

High

• Category: analysis-evasion
• Confidence: 100% confidence

Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 29)May include surrounding context.

md

- One folder per skill. The main file is always `SKILL.md` (exact name, case-sensitive).

Ae1

High

• Category: analysis-evasion
• Confidence: 100% confidence

Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 90)May include surrounding context.

md

- One folder per skill. The main file is always `SKILL.md` (exact name, case-sensitive).

Self-Modification

High

• Category: Rogue Agent
• Confidence: 94% confidence

Finding

The process explicitly instructs the agent to write or update SKILL.md and supporting files. This operationalizes self-modification rather than merely describing it, creating a direct pathway for persistent changes to the agent environment if the task is triggered by malicious content or broad matching.

Content

Scanner excerpt · SKILL.md (reported line 107)May include surrounding context.

md

1. Read the target skill path and any source.
2. Apply the checklist and structure above. Don't change behavior unless the user asked.
3. Write or update SKILL.md (and supporting files if needed). Only document what the skill does; don't add capabilities that aren't there.

## Examples

Vague Triggers

Medium

• Category: Not specified by scanner
• Confidence: 94% confidence

Finding

The description uses very broad trigger phrases such as 'create skill' and 'write skill' without narrowing scope or exclusion criteria. In a host that auto-selects skills by natural-language matching, this can cause the skill to activate in unrelated contexts and steer the assistant into editing skill files unexpectedly, increasing the chance of unsafe or unintended self-modification workflows.

Content

No source excerpt is available for this finding.

Whitespace Padding

Medium

• Category: Prompt Injection
• Confidence: 70% confidence

Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 45)May include surrounding context.

md

**name** and **description** are required. **name** must match the folder name (kebab-case). **description** should follow: **`[What it does] + [When to use it] + [Key capabilities]`** — under 1024 characters; no XML angle brackets (`<` `>`). Avoid vague lines like "Helps with projects"; include specific trigger phrases and, if relevant, file types (e.g. ".fig files", "PDF contract review").

| Field                      | Required    | Description                                                                                                                                           |
| :------------------------- | :---------- | :---------------------------------------------------------------------------------------------------------------------------------------------------- |
| `name`                     | Yes         | kebab-case only; must match skill folder name. Lowercase letters, numbers, and hyphens (max 64 characters).                                           |
| `description`              | Yes         | **`[What it does] + [When to use it] + [Key capabilities]`**. Include trigger phrases (e.g. "Use when user says save, /save"). Mention file types if relevant. Under 1024 characters. No `<` or `>`.        |

Autonomous Decision Making

Medium

• Category: Excessive Agency
• Confidence: 75% confidence

Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 52)May include surrounding context.

md

| `argument-hint`            | No          | Hint shown during autocomplete to indicate expected arguments. Example: `[issue-number]` or `[filename] [format]`.                                    |
| `disable-model-invocation` | No          | Set to `true` to disable automatic loading; use for workflows you want to trigger manually with `/name`. Default: `false`. |
| `user-invocable`           | No          | Set to `false` to hide from the `/` menu. Use for background knowledge users shouldn't invoke directly. Default: `true`.                              |
| `allowed-tools`            | No          | Tools the assistant can use without asking permission when this skill is active.                                                                             |
| `model`                    | No          | Model to use when this skill is active.                                                                                                               |
| `context`                  | No          | Set to `fork` to run in a forked subagent context.                                                                                                    |
| `agent`                    | No          | Which subagent type to use when `context: fork` is set.                                                                                               |

Whitespace Padding

Medium

• Category: Prompt Injection
• Confidence: 70% confidence

Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 53)May include surrounding context.

md

| `disable-model-invocation` | No          | Set to `true` to disable automatic loading; use for workflows you want to trigger manually with `/name`. Default: `false`. |
| `user-invocable`           | No          | Set to `false` to hide from the `/` menu. Use for background knowledge users shouldn't invoke directly. Default: `true`.                              |
| `allowed-tools`            | No          | Tools the assistant can use without asking permission when this skill is active.                                                                             |
| `model`                    | No          | Model to use when this skill is active.                                                                                                               |
| `context`                  | No          | Set to `fork` to run in a forked subagent context.                                                                                                    |
| `agent`                    | No          | Which subagent type to use when `context: fork` is set.                                                                                               |
| `hooks`                    | No          | Hooks scoped to this skill's lifecycle. See [Hooks in skills and agents](/en/hooks#hooks-in-skills-and-agents) for configuration format.              |

Whitespace Padding

Medium

• Category: Prompt Injection
• Confidence: 70% confidence

Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 54)May include surrounding context.

md

| `user-invocable`           | No          | Set to `false` to hide from the `/` menu. Use for background knowledge users shouldn't invoke directly. Default: `true`.                              |
| `allowed-tools`            | No          | Tools the assistant can use without asking permission when this skill is active.                                                                             |
| `model`                    | No          | Model to use when this skill is active.                                                                                                               |
| `context`                  | No          | Set to `fork` to run in a forked subagent context.                                                                                                    |
| `agent`                    | No          | Which subagent type to use when `context: fork` is set.                                                                                               |
| `hooks`                    | No          | Hooks scoped to this skill's lifecycle. See [Hooks in skills and agents](/en/hooks#hooks-in-skills-and-agents) for configuration format.              |

Whitespace Padding

Medium

• Category: Prompt Injection
• Confidence: 70% confidence

Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL.md (reported line 55)May include surrounding context.

md

| `allowed-tools`            | No          | Tools the assistant can use without asking permission when this skill is active.                                                                             |
| `model`                    | No          | Model to use when this skill is active.                                                                                                               |
| `context`                  | No          | Set to `fork` to run in a forked subagent context.                                                                                                    |
| `agent`                    | No          | Which subagent type to use when `context: fork` is set.                                                                                               |
| `hooks`                    | No          | Hooks scoped to this skill's lifecycle. See [Hooks in skills and agents](/en/hooks#hooks-in-skills-and-agents) for configuration format.              |

**Optional:** `license` (e.g. MIT for open source), `compatibility` (1–500 chars: product, system deps, network), `metadata` (e.g. author, version, mcp-server). See [complete skills guide](https://resources.anthropic.com/hubfs/The-Complete-Guide-to-Building-Skill-for-Claude.pdf).

Skill Enumeration

Medium

• Category: Agent Snooping
• Confidence: 80% confidence

Finding

Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Content

Scanner excerpt · SKILL.md (reported line 119)May include surrounding context.

md

**User says:** "Update the save skill to mention update-gitignore."

**Actions:** Open the save skill's SKILL.md (e.g. `skills/save/SKILL.md`). Add the new step or reference in Process. Update description if trigger phrases change. Sync coordinator and [references/coordinator-flows.md](../../agents/references/coordinator-flows.md) if the Save flow changes.

## Troubleshooting

Static analysis

No suspicious patterns detected.