Last30days Zh

AdvisoryAudited by Static analysis on May 8, 2026.

Overview

No suspicious patterns detected.

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.

NoteHigh Confidence

ASI05: Unexpected Code Execution

What this means

Installing and using the skill lets the agent run the bundled research scripts on your machine.

Why it was flagged

The skill is operated by running bundled Bash/Python scripts. This is central to the research CLI workflow and is disclosed, but users should understand they are executing local code.

Skill content

bash scripts/run-last30days.sh "$ARGUMENTS" --emit=compact
python3 scripts/last30days.py "$ARGUMENTS" --api-key="$AISA_API_KEY"

Recommendation

Review the scripts if you are in a sensitive environment, and run the skill from a trusted copy of the package.

NoteHigh Confidence

ASI03: Identity and Privilege Abuse

What this means

The skill can use your AISA account quota/permissions to perform hosted searches and model calls.

Why it was flagged

The code sends the disclosed AISA API key as a bearer token to AISA-hosted endpoints. This is expected for the AISA-backed search and synthesis features.

Skill content

AISA_BASE_URL = "https://api.aisa.one"
...
"Authorization": f"Bearer {api_key}"

Recommendation

Use a dedicated or least-privileged AISA key if available, and avoid sharing the key in transcripts or saved output.

NoteHigh Confidence

ASI07: Insecure Inter-Agent Communication

What this means

Your research topic and related query/context data may be sent to AISA and relevant hosted search providers.

Why it was flagged

The skill discloses that research planning, reranking, synthesis, web search, and some platform searches are routed through AISA-hosted services.

Skill content

通过 AISA 提供规划、重排、综合、grounded web search、X/Twitter、YouTube 和 Polymarket。

Recommendation

Do not use the skill for highly confidential research topics unless you are comfortable with the AISA/provider data flow.

NoteHigh Confidence

ASI06: Memory and Context Poisoning

What this means

A local config file may contain credentials or provider settings that affect future runs from that directory.

Why it was flagged

The skill can read a local repo/current-directory config file for API keys and options, and warns if that file is too broadly readable. This is disclosed and limited to configuration, not persistent agent memory.

Skill content

CONFIG_DIR = Path.cwd() / ".last30days-data"
CONFIG_FILE = CONFIG_DIR / "config.env"
...
_check_file_permissions(path)

Recommendation

Keep `.last30days-data/config.env` private, set restrictive permissions such as `chmod 600`, and do not place untrusted configuration files in working directories.