Trust Assessment
openclaw-news received a trust score of 10/100, placing it in the Untrusted category. This skill has significant security findings that require attention before use in production.
SkillShield's automated analysis identified 10 findings: 3 critical, 4 high, 2 medium, and 1 low severity. Key findings include Arbitrary command execution, Missing required field: name, Dangerous call: subprocess.run().
The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. The LLM Behavioral Safety layer scored lowest at 31/100, indicating areas for improvement.
Last analyzed on February 14, 2026 (commit 13146e6a). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings10
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| CRITICAL | Arbitrary command execution Python shell execution (os.system, subprocess) Review all shell execution calls. Ensure commands are static (not built from user input), use absolute paths, and are strictly necessary. Prefer library APIs over shell commands. | Manifest | skills/arc-claw-bot/openclaw-news/scripts/agent_collect.py:54 | |
| CRITICAL | Arbitrary command execution Python shell execution (os.system, subprocess) Review all shell execution calls. Ensure commands are static (not built from user input), use absolute paths, and are strictly necessary. Prefer library APIs over shell commands. | Manifest | skills/arc-claw-bot/openclaw-news/scripts/agent_collect.py:98 | |
| CRITICAL | Skill instructs LLM to execute scripts via cron prompt The `SKILL.md` defines cron jobs with a `--prompt` argument that directly instructs the host LLM to "Execute scripts/collect_news.sh from skills/openclaw-news/, then format and deliver the briefing." As the skill package itself is untrusted content, this constitutes a direct prompt injection attempt, instructing the LLM to perform actions that could bypass intended safety mechanisms. The LLM orchestration layer should validate or sanitize prompts originating from untrusted skill definitions. If the skill is untrusted, the LLM should not directly execute arbitrary instructions from its prompt. Instead, it should invoke a predefined tool or function for skill execution with strictly controlled parameters. | LLM | SKILL.md:58 | |
| HIGH | Dangerous call: subprocess.run() Call to 'subprocess.run()' detected in function 'run_gh'. This can execute arbitrary code. Avoid using dangerous functions like exec/eval/os.system. Use safer alternatives. | Static | skills/arc-claw-bot/openclaw-news/scripts/agent_collect.py:54 | |
| HIGH | Dangerous call: subprocess.run() Call to 'subprocess.run()' detected in function 'collect_clawhub'. This can execute arbitrary code. Avoid using dangerous functions like exec/eval/os.system. Use safer alternatives. | Static | skills/arc-claw-bot/openclaw-news/scripts/agent_collect.py:98 | |
| HIGH | Direct execution of shell scripts and external commands with interpolated variables The skill's primary function involves executing shell scripts (`collect_news.sh`, `format_briefing.sh`) and external binaries (`gh`, `clawdhub`, `jq`, `curl`) via `subprocess.run` or direct shell calls. The pattern of interpolating shell variables directly into `python3 -c "..."` commands (e.g., in `collect_news.sh` functions `merge_field` and `add_error`) or `gh api --jq` arguments presents a significant command injection risk. If any of the interpolated variables were to become controllable by untrusted input, it could lead to arbitrary command execution or data manipulation. Avoid interpolating shell variables directly into `python3 -c` commands. Instead, pass data via stdin/stdout or environment variables, or use a dedicated Python script that takes arguments securely. For `gh api --jq`, ensure all interpolated values are strictly validated or escaped if they originate from untrusted sources. | LLM | scripts/collect_news.sh:80 | |
| HIGH | API key used directly in `curl` command in shell script The `collect_news.sh` script uses the `$MOLTBOOK_API_KEY` environment variable directly within a `curl` command. This poses a credential harvesting risk, as the API key could be exposed in process lists, logs, or exfiltrated to an attacker-controlled endpoint if the `$MOLTBOOK_API_URL` were compromised or manipulated. Implement secure credential management practices. Avoid passing API keys directly as environment variables or command-line arguments. If a Python script is used, consider using a library like `requests` and fetching the key from a secure vault at runtime. Ensure the API endpoint is fixed and not configurable by untrusted input. | LLM | scripts/collect_news.sh:250 | |
| MEDIUM | Missing required field: name The 'name' field is required for claude_code skills but is missing from frontmatter. Add a 'name' field to the SKILL.md frontmatter. | Static | skills/arc-claw-bot/openclaw-news/SKILL.md:1 | |
| MEDIUM | Broad access granted through `gh` CLI and web search tool The skill requires the `gh` CLI to be installed and authenticated, and it uses `gh api` to query GitHub. This grants the skill potentially broad read access to GitHub repositories, issues, and pull requests, depending on the scope of the authenticated `gh` token. Furthermore, the skill relies on the agent's `web_search` capabilities (e.g., Brave Search), which can perform arbitrary web queries, potentially leading to information gathering beyond the intended scope if misused. Implement least privilege for `gh` CLI tokens, ensuring they only have the minimum necessary read permissions. For web search, consider sandboxing the search tool or restricting its domains/query types if possible, or ensure the agent's execution environment strictly limits what the skill can do with search results. | LLM | SKILL.md:35 | |
| LOW | Unpinned external CLI dependencies The skill relies on external command-line tools (`gh`, `clawdhub`, `jq`) without specifying or checking their versions. This introduces a supply chain risk, as updates to these tools could introduce breaking changes, vulnerabilities, or unexpected behavior that the skill is not designed to handle, potentially leading to instability or security issues. Pin specific versions for all external dependencies and include version checks in the skill's setup or execution logic. Provide instructions for how to install specific versions of these tools. | LLM | SKILL.md:35 |
Scan History
Embed Code
[](https://skillshield.io/report/bc33704c170a575d)
Powered by SkillShield