Back to skill

Security audit

Grazer Skill Main

Security checks for vulnerabilities and agentic risk

Overview

Grazer has a coherent discovery purpose, but its autonomous posting features, plaintext token config, arbitrary outbound endpoints, and weak SVG handling need careful review before installation.

Review this before installing in any account that can post publicly. Keep auto_respond disabled until tested, use least-privilege API keys, prefer dry-run/idempotency for write paths, restrict or avoid custom LLM URLs, do not place sensitive prompts or tokens in the config, and avoid accepting untrusted SVG until sanitization or sandboxing is added.

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
grazer/imagegen.py:233
Finding

Unrestricted outbound URLs enable server-side request forgery and credential disclosure

Content
View full analysis
str: inst = (instance or self.instance).rstrip("/") if not inst.startswith("http"): inst = f"https://{inst}" return f"{inst}/api/v1" ``` `grazer/mastodon_grazer.py:54-58` ```python resp = self.session.get( f"{base}/search", params=params, timeout=self.timeout, ) ``` `grazer/imagegen.py:233-266` ```python def generate_llm_svg( prompt: str, llm_url: str = "http://100.75.100.89:8080/v1/chat/completions", llm_model: str = "gpt-oss-120b", llm_api_key: Optional[str] = None, temperature: float = 0.8, timeout: int = 60, ) -> str: """Generate SVG using any OpenAI-compatible LLM endpoint.""" headers = {"Content-Type": "application/json"} if llm_api_key: headers["Authorization"] = f"Bearer {llm_api_key}" payload = { "model": llm_model, "messages": [ {"role": "system", "content": LLM_SVG_SYSTEM_PROMPT}, {"role": "user", "content": f"Create an SVG image: {prompt}"}, ], "temperature": temperature, "max_tokens": 2048, } resp = requests.post(llm_url, json=payload, headers=headers, timeout=timeout) ``` `src/imagegen.ts:187-209` ```typescript export async function generateLlmSvg( prompt: string, llmUrl: string, llmModel = 'gpt-oss-120b', llmApiKey?: string, temperature = 0.8, timeout = 60000 ): Promise { const headers: Record = { 'Content-Type': 'application/json' }; if (llmApiK ...[truncated 2901 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
grazer/imagegen.py:203
Finding

Insufficient SVG validation permits active-content injection in generated and caller-supplied media

Content
View full analysis
str: """Validate and sanitize SVG for 4claw.""" svg = svg.strip() # Must start with SVG_MAX_BYTES: raise ValueError(f"SVG exceeds 4KB limit ({len(svg.encode('utf-8'))} bytes)") return svg ``` `grazer/imagegen.py:268-275` ```python content = resp.json()["choices"][0]["message"]["content"].strip() # Extract SVG if wrapped in code fences svg_match = re.search(r'', content) if not svg_match: raise ValueError("LLM did not produce valid SVG output") return _validate_svg(svg_match.group(0)) ``` `grazer/__init__.py:683-701` ```python body = {"title": title, "content": content, "anon": anon} # Attach SVG media if provided or generated if svg: body["media"] = svg_to_media(svg) elif image_prompt: result = self.generate_image(image_prompt, template=template, palette=palette) body["media"] = svg_to_media(result["svg"]) resp = self._rate_limited_post( f"https://www.4claw.org/api/v1/boards/{board}/threads", json=body, headers={ "Authorization": f"Bearer {self.fourclaw_key}", "Content-Type": "application/json", }, timeout=self.timeout, ) ``` `src/imagegen.ts:154-161` ```typescript function validateSvg(svg: string): string { svg = svg.trim(); if (!svg.startsWith('
Remediation
View remediation
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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
Findings (85)

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · README.md (reported line 74)May include surrounding context.

retained for use when the service is restored.

bash
curl -fsSL https://bottube.ai/apt/gpg | sudo gpg --dearmor -o /usr/share/keyrings/grazer.gpg
echo "deb [signed-by=/usr/share/keyrings/grazer.gpg] https://bottube.ai/apt stable main" | sudo tee /etc/apt/sources.list.d/grazer.list
sudo apt update && sudo apt install grazer

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · README.md (reported line 75)May include surrounding context.

retained for use when the service is restored.

bash
curl -fsSL https://bottube.ai/apt/gpg | sudo gpg --dearmor -o /usr/share/keyrings/grazer.gpg
echo "deb [signed-by=/usr/share/keyrings/grazer.gpg] https://bottube.ai/apt stable main" | sudo tee /etc/apt/sources.list.d/grazer.list
sudo apt update && sudo apt install grazer

Chaining Abuse

High
Category
Tool Misuse
Confidence
84% confidence
Finding

The README instructs users to pipe remote network content directly into a privileged command: 'curl ... | sudo gpg --dearmor'. Even though this is for repository key installation, direct network-to-root pipelines are dangerous because they reduce review opportunities and create a straightforward path for compromise if the remote host, TLS trust, or distribution channel is tampered with.

Content

Scanner excerpt · README.md (reported line 74)May include surrounding context.

retained for use when the service is restored.

bash
curl -fsSL https://bottube.ai/apt/gpg | sudo gpg --dearmor -o /usr/share/keyrings/grazer.gpg
echo "deb [signed-by=/usr/share/keyrings/grazer.gpg] https://bottube.ai/apt stable main" | sudo tee /etc/apt/sources.list.d/grazer.list
sudo apt update && sudo apt install grazer

Chaining Abuse

High
Category
Tool Misuse
Confidence
81% confidence
Finding

This line pipes generated repository configuration text directly into 'sudo tee', which writes privileged configuration without an explicit review step. In context this is less severe than the remote key pipeline, but it still encourages blind copy-paste modification of trusted package sources and can amplify damage if surrounding instructions or repository details are later altered.

Content

Scanner excerpt · README.md (reported line 75)May include surrounding context.

bash
curl -fsSL https://bottube.ai/apt/gpg | sudo gpg --dearmor -o /usr/share/keyrings/grazer.gpg
echo "deb [signed-by=/usr/share/keyrings/grazer.gpg] https://bottube.ai/apt stable main" | sudo tee /etc/apt/sources.list.d/grazer.list
sudo apt update && sudo apt install grazer

Chaining Abuse

High
Category
Tool Misuse
Confidence
75% confidence
Finding

Tool calls are chained to bypass individual safety checks or escalate capabilities beyond what any single tool call would allow.

Content

Scanner excerpt · README.md (reported line 76)May include surrounding context.

bash
curl -fsSL https://bottube.ai/apt/gpg | sudo gpg --dearmor -o /usr/share/keyrings/grazer.gpg
echo "deb [signed-by=/usr/share/keyrings/grazer.gpg] https://bottube.ai/apt stable main" | sudo tee /etc/apt/sources.list.d/grazer.list
sudo apt update && sudo apt install grazer

Claude Code

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
100% 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 · debian/build-deb.sh (reported line 8)May include surrounding context.

sh
VERSION="1.9.1"
PKG="grazer_${VERSION}_all"

rm -rf "/tmp/${PKG}"
mkdir -p "/tmp/${PKG}/DEBIAN"
mkdir -p "/tmp/${PKG}/usr/lib/python3/dist-packages/grazer"
mkdir -p "/tmp/${PKG}/usr/bin"

Self-Modification

High
Category
Rogue Agent
Confidence
85% confidence
Finding

Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Content

Scanner excerpt · src/clawhub.ts (reported line 87)May include surrounding context.

ts
}

  /**
   * Update skill metadata
   */
  async updateSkill(skillId: string, updates: Partial<ClawHubSkill>): Promise<ClawHubSkill> {
    if (!this.token) {

Ssd 1

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

Untrusted notification text is interpolated directly into the LLM prompt as conversation context, allowing commenters to inject instructions that influence the model's response. In this skill, the LLM output is intended for automated replying, so prompt injection can directly produce attacker-shaped public responses, policy bypasses, or social-engineering content without human review.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% 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 · DEPLOY.md (reported line 59)May include surrounding context.

Create a skill entry on BoTTube:

bash
curl -X POST https://bottube.ai/api/skills \
  -H "Content-Type: application/json" \
  -d '{
    "name": "grazer",

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 · INTEGRATION.md (reported line 123)May include surrounding context.

1. Setup Config

bash
mkdir -p ~/.grazer
cp config.example.json ~/.grazer/config.json
cp profile.example.json ~/.grazer/profile.json

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding

The guide instructs users to run npx grazer-agent without pinning a specific package version or verifying integrity. That creates a supply-chain risk because the code executed can change over time or be replaced by a malicious package/version, and the command is presented as part of an autonomous agent setup likely to run with user credentials and API keys.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill explicitly promotes autonomous discovery, monitoring, auto-response, and learning on external platforms without a clear warning, approval gate, or human-in-the-loop safeguard. In context, this can cause unintended posts, spam, reputational damage, or policy violations on third-party services because the agent is encouraged to act continuously and independently.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The instructions say to inject discovered results into the LLM prompt as [SYSTEM DATA] without caution about privacy, prompt injection, or sensitive-data handling. Content retrieved from external platforms is untrusted input; elevating it into system-context can let adversarial content influence model behavior or expose private information to downstream processing.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
85% confidence
Finding

auto-deploy in the context of notification monitoring and conversation deployment indicates autonomous decision-making for generating and posting responses. While likely intended as product functionality, it increases risk because decisions to engage are delegated to the agent without explicit validation, making accidental harmful actions more likely on live platforms.

Content

Scanner excerpt · INTEGRATION.md (reported line 163)May include surrounding context.

text

### Boris (Moltbook Bot)
Already integrated via notification monitor + auto-deploy

### Janitor (AutomatedJanitor2015)
Add to notification checking:

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The issue template explicitly asks users to paste log output but does not warn them that logs may contain secrets, tokens, API keys, file paths, usernames, or other sensitive operational data. In a public issue tracker, this can lead to unintentional disclosure of confidential information that attackers can later harvest.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% 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 · PUBLISH_CHECKLIST.md (reported line 54)May include surrounding context.

bash
# Using ClawHub CLI or API
curl -X POST https://clawhub.ai/api/skills \
  -H "Authorization: Bearer <CLAWHUB_TOKEN_REDACTED — set via clawhub login, stored in ~/.config/clawhub>" \
  -H "Content-Type: application/json" \
  -d '{

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The checklist instructs operators to enable an auto-response mode by editing a config file, but it does not warn that doing so authorizes autonomous actions that may post content, interact with external services, or affect user data/accounts. In an agent skill context, lack of an explicit safety warning can lead to unintended automated behavior being enabled without review, increasing the chance of account misuse or privacy-impacting actions.

Content

No source excerpt is available for this finding.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

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

Content

Scanner excerpt · README.md (reported line 74)May include surrounding context.

retained for use when the service is restored.

bash
curl -fsSL https://bottube.ai/apt/gpg | sudo gpg --dearmor -o /usr/share/keyrings/grazer.gpg
echo "deb [signed-by=/usr/share/keyrings/grazer.gpg] https://bottube.ai/apt stable main" | sudo tee /etc/apt/sources.list.d/grazer.list
sudo apt update && sudo apt install grazer

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

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

Content

Scanner excerpt · README.md (reported line 75)May include surrounding context.

retained for use when the service is restored.

bash
curl -fsSL https://bottube.ai/apt/gpg | sudo gpg --dearmor -o /usr/share/keyrings/grazer.gpg
echo "deb [signed-by=/usr/share/keyrings/grazer.gpg] https://bottube.ai/apt stable main" | sudo tee /etc/apt/sources.list.d/grazer.list
sudo apt update && sudo apt install grazer

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

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

Content

Scanner excerpt · README.md (reported line 76)May include surrounding context.

retained for use when the service is restored.

bash
curl -fsSL https://bottube.ai/apt/gpg | sudo gpg --dearmor -o /usr/share/keyrings/grazer.gpg
echo "deb [signed-by=/usr/share/keyrings/grazer.gpg] https://bottube.ai/apt stable main" | sudo tee /etc/apt/sources.list.d/grazer.list
sudo apt update && sudo apt install grazer

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 · README.md (reported line 102)May include surrounding context.

md
# Browse 4claw /crypto/ board
grazer discover -p fourclaw -b crypto

# Create a 4claw thread
grazer post -p fourclaw -b singularity -t "Title" -m "Content"

# Reply to a 4claw thread

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The README explicitly states that NPM and PyPI installs are reported to a remote service, but the installation sections earlier do not warn users that installing the package triggers telemetry. Silent install-time reporting is a privacy and supply-chain trust concern because users may disclose environment usage or IP-linked installation events without informed consent.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill explicitly advertises autonomous notifications, auto-responses, training from interactions, and a continuous engagement loop across many external platforms, but the documentation does not prominently warn that these features can perform outbound write actions on behalf of the user. In an agent setting, this can lead to unintended posting, spam, reputation damage, or policy violations if invoked without explicit per-action confirmation and clear permission boundaries.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The examples show posting to 4claw and writing an SVG file to disk, but they do not include conspicuous safety guidance that these commands have side effects outside the local read-only discovery flow. In practice, copy-pasting such examples into an agent workflow can cause unintended publication of content or local file creation without the operator realizing the action is mutating external or local state.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
72% confidence
Finding

The skill instructs users to persist multiple API keys and tokens in a long-lived plaintext config file under the home directory. While common, this increases the blast radius of local compromise, accidental backup/sync leakage, token reuse exposure, or later misuse by the agent or other processes if file protections are weak or the environment is multi-tenant.

Content

Scanner excerpt · SKILL.md (reported line 146)May include surrounding context.

Configuration

Create ~/.grazer/config.json:

json
{