News Aggregator Skill

PassAudited by VirusTotal on May 12, 2026.

Overview

Type: OpenClaw Skill Name: news-aggregator-skill-4 Version: 0.1.0 The `SKILL.md` file instructs the AI agent to automatically expand user-provided keywords (e.g., 'AI' to 'AI,LLM,GPT,...') before passing them to the `fetch_news.py` script. This is a form of prompt injection where the agent modifies user input based on internal instructions. While the immediate objective of this keyword expansion is benign (improving search results), it demonstrates a risky capability for the agent to alter user input, which could be a vector for injecting unintended arguments or commands under different instructions. Additionally, the skill involves downloading and parsing content from external websites via the `--deep` flag, which is a high-risk operation, though necessary for its stated purpose.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

The skill may not run as described, or an agent/user may need to locate or provide code that was not included in the package.

Why it was flagged

The skill references helper/template files, but the supplied manifest contains only SKILL.md and no install spec, so the runnable implementation and template provenance are not present in the reviewed artifacts.

Skill content
python3 scripts/fetch_news.py --source all --limit 15 --deep ... READ the content of `templates.md` in the skill directory
Recommendation

Before executing, verify that scripts/fetch_news.py and templates.md are present, expected, and from a trusted source.

NoteHigh Confidence
ASI01: Agent Goal Hijack
What this means

A report may include supplementary older stories even when the user asked for a strict recent time range.

Why it was flagged

For sparse time-window queries, the skill tells the agent to include older items beyond the user's requested window, although it also requires those items to be clearly annotated.

Skill content
If the list is short, you MUST include high-value/high-heat items from a wider range (e.g. past 24h)
Recommendation

If strict filtering matters, ask explicitly for only items within the requested time window and review any annotations on older items.

What this means

Local report files may remain after use and could reveal the news topics or briefings the user requested.

Why it was flagged

The skill creates persistent local report files containing generated summaries and links; this is scoped and purpose-aligned, but it leaves durable output on disk.

Skill content
Always save the full report to `reports/` directory with a timestamped filename
Recommendation

Store reports only where appropriate and delete old reports if the topics or generated briefings are sensitive.