Back to skill

Security audit

Writing Coach Pro

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent writing coach, but it automatically retains and syncs writing-derived profile/history data in ways users should review before installing.

Review this before installing if you handle confidential writing. Expect local profile/history files, possible stored excerpts, automatic learning, and dashboard sync when the dashboard is present; disable those features if the host supports it, and be careful with the documented reset command because it permanently deletes the skill's stored data.

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
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Rogue AgentSelf-Modification, Session Persistence
Findings (29)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SECURITY.md (reported line 37)May include surrounding context.

md
- Review your LLM provider's data handling policies if you process sensitive documents
- Use a local model (Ollama, llama.cpp) if you need text to stay on your machine
- The `data/` directory contains analysis metadata — protect it like any personal file
- Run `rm -rf ~/.openclaw/skills/writing-coach-pro/data` to delete all stored analysis data

## Reporting Issues

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
85% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SECURITY.md (reported line 37)May include surrounding context.

md
- Review your LLM provider's data handling policies if you process sensitive documents
- Use a local model (Ollama, llama.cpp) if you need text to stay on your machine
- The `data/` directory contains analysis metadata — protect it like any personal file
- Run `rm -rf ~/.openclaw/skills/writing-coach-pro/data` to delete all stored analysis data

## Reporting Issues

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SECURITY.md (reported line 37)May include surrounding context.

md
- Review your LLM provider's data handling policies if you process sensitive documents
- Use a local model (Ollama, llama.cpp) if you need text to stay on your machine
- The `data/` directory contains analysis metadata — protect it like any personal file
- Run `rm -rf ~/.openclaw/skills/writing-coach-pro/data` to delete all stored analysis data

## Reporting Issues

Missing User Warnings

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

The guide states that the product "gets smarter every time you use it" and tracks accepted/rejected suggestions, which implies behavioral data collection without any visible consent, retention, or privacy explanation. In a writing tool, usage patterns and submitted text may contain sensitive personal or business content, so silent profiling creates a meaningful privacy risk.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The guide says Writing Coach Pro automatically feeds writing stats to the NormieClaw Dashboard, but it does not present this as a privacy-sensitive external sharing action or offer consent controls. Automatic transmission of writing-related analytics to another component can expose sensitive work habits, topics, or metadata, especially if users do not realize a separate dashboard integration is receiving data.

Content

No source excerpt is available for this finding.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

The command rm -rf ~/.openclaw/skills/writing-coach-pro/data is a destructive filesystem operation presented for manual execution without safeguards, confirmations, or validation. In context it is intended to reset the skill, not to attack the user, but destructive shell commands in end-user documentation are hazardous because path expansion mistakes, copy/paste errors, or modified environments can lead to unintended deletion.

Content

Scanner excerpt · SETUP-PROMPT.md (reported line 90)May include surrounding context.

Want to start fresh Delete the data folder and re-run setup:

bash
rm -rf ~/.openclaw/skills/writing-coach-pro/data
bash ~/.openclaw/skills/writing-coach-pro/scripts/setup.sh

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

The command rm -rf ~/.openclaw/skills/writing-coach-pro/data is a destructive filesystem operation presented for manual execution without safeguards, confirmations, or validation. In context it is intended to reset the skill, not to attack the user, but destructive shell commands in end-user documentation are hazardous because path expansion mistakes, copy/paste errors, or modified environments can lead to unintended deletion.

Content

Scanner excerpt · SETUP-PROMPT.md (reported line 90)May include surrounding context.

Want to start fresh Delete the data folder and re-run setup:

bash
rm -rf ~/.openclaw/skills/writing-coach-pro/data
bash ~/.openclaw/skills/writing-coach-pro/scripts/setup.sh

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The README advertises "Persistent Style Memory" that learns user preferences over time, but it does not explain what data is stored, where it is stored, how long it is retained, or whether sensitive text may be persisted. For a writing assistant handling potentially confidential drafts, undisclosed persistence creates privacy and data-governance risk and may lead users to share content they would not otherwise store.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The trigger phrase "review this" is extremely generic and likely to appear in normal conversations, causing the skill to activate unintentionally. In an agent context, broad activation can expose user text to processing, invoke rewriting behavior at the wrong time, or interfere with unrelated workflows, increasing privacy and integrity risk.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The setup guide suggests generic activation phrases such as "review this," "rewrite this," and "coach me on this." In a markdown file, these are ambiguous and overlap with common everyday speech, with no negative examples or scope limits explaining when the skill should or should not activate.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The reset instructions tell users to delete the data directory with rm -rf but do not clearly warn that this permanently removes learned preferences and history. This is dangerous because users may execute the command without understanding that the loss is irreversible and affects accumulated writing-profile data.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
82% confidence
Finding

The skill is explicitly designed around persistent preferences, trends, and longitudinal tracking, which creates session persistence risk. While persistence is part of the feature set, it becomes security-relevant because user writing often contains sensitive information and the skill normalizes saving derived data across sessions.

Content

Scanner excerpt · SKILL.md (reported line 14)May include surrounding context.

md
Use when the user asks to:
- Review, edit, proofread, or polish writing
- Check grammar, style, or tone
- Rewrite or rephrase text
- Score readability or grade level
- Check document consistency
- Set writing preferences or style guide

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill states that profiles persist across sessions and evolve over time, but it does not present a clear user warning or consent mechanism before storing writing preferences and inferred behavior. Because writing submissions can contain sensitive personal, business, or legal content, silent retention creates meaningful privacy and compliance risk.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The implicit learning loop updates the user's profile based on acceptance and rejection patterns without explicit opt-in. Behavioral profiling from interaction history can reveal preferences, habits, and potentially sensitive contexts, and doing this automatically is risky in an agent that processes user-authored text.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 244)May include surrounding context.

md
Always end with "Top 3 Priorities" — the user's attention is finite. Surface what matters most.

### Rewrite Mode

**Trigger**: "rewrite this", "polish this", "make this better", "clean this up".

Ssd 3

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill combines persistent logging, profile history, session statistics, and dashboard reuse of user writing-derived metadata across multiple files and tables. This broad retention exceeds the immediate task of proofreading and raises the risk of unintended disclosure, secondary use, or access by other local components.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The learning log stores dated session history, suggestion outcomes, and context labels, which creates a persistent behavioral record without a clear warning or retention policy. Such logs can expose patterns about a user's work, communication style, and activities beyond the immediate writing task.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The dashboard sync exports session and issue data into additional tables without clearly warning the user that analysis results and profile history will be retained in another subsystem. Replicating user-derived data increases exposure surface and may make deletion or access control harder.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill exposes shell-script execution via export-report.sh and style-check.sh even though the document does not define strict input validation, execution constraints, or a narrow trust boundary for those scripts. In an agent setting, adding command execution expands the attack surface from text analysis into local code execution, which could be abused through crafted filenames, argument injection, or unsafe downstream script behavior.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

This markdown spec instructs users to delete local logs and run SQL DELETE commands that remove dashboard records, but it does not include any warning that these actions are destructive and may permanently erase historical session data. Under the markdown-file criteria for missing user warnings, behaviors affecting user data should be explicitly disclosed.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The manifest describes the skill only as "Writing analysis trends, style profile, and improvement tracking," which is a broad capability description rather than a narrowly scoped invocation condition. In a manifest file, this lack of specificity can lead to ambiguous matching against general writing-related requests, with no explicit exclusions or trigger constraints to reduce unintended activation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The script writes a persistent report containing session-derived writing history and statistics to disk under the user's home directory without any consent prompt, privacy warning, retention control, or file permission hardening. In a writing-coach context, exported content can reveal sensitive drafts, habits, dates, and performance data, increasing exposure to local compromise, backup leakage, or unintended sharing.

Content

No source excerpt is available for this finding.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
80% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · scripts/setup.sh (reported line 31)May include surrounding context.

sh
mkdir -p "${REPORTS_DIR}"

# Set permissions — owner only
chmod 700 "${DATA_DIR}"
chmod 700 "${SESSION_DIR}"
chmod 700 "${REPORTS_DIR}"

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
80% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · scripts/setup.sh (reported line 32)May include surrounding context.

sh
mkdir -p "${REPORTS_DIR}"

# Set permissions — owner only
chmod 700 "${DATA_DIR}"
chmod 700 "${SESSION_DIR}"
chmod 700 "${REPORTS_DIR}"

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
80% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · scripts/setup.sh (reported line 33)May include surrounding context.

sh
mkdir -p "${REPORTS_DIR}"

# Set permissions — owner only
chmod 700 "${DATA_DIR}"
chmod 700 "${SESSION_DIR}"
chmod 700 "${REPORTS_DIR}"