v1.0.10

Odaily Skill

ReviewClawScan verdict for this skill. Analyzed May 1, 2026, 8:11 AM.

Analysis

The skill mostly matches its crypto-market purpose, but it needs review because normal use runs local shell commands that quietly install Python packages and execute local code, with optional credential-backed persistence.

GuidanceInstall only if you are comfortable with this skill running local Python via Bash. Prefer using a virtual environment, pinning dependencies, removing the automatic pip install from normal use, and checking whether SUPABASE_URL, SUPABASE_KEY, or COINGECKO_API_KEY are set in your environment. Treat the output as market information, not investment advice.

Findings (5)

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.

Abnormal behavior control

Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.

Unexpected Code Execution

SeverityMediumConfidenceHighStatusConcern

SKILL.md

SKILL_DIR=$(find ~/.openclaw ~/.claude -name "run.py" -path "*/odai*" ... ) && cd "$SKILL_DIR" && pip install -r requirements.txt -q 2>/dev/null && python3 run.py <工具名> '<JSON参数>'

Normal tool invocation includes a broad local search, quiet dependency installation, and Python execution rather than a one-time reviewed setup step.

User impactUsing the skill can modify the local Python environment and execute whichever matching local run.py is found.

RecommendationInstall dependencies once in a dedicated virtual environment, remove runtime pip install from the tool command, use a fixed skill path, and avoid suppressing install output.

Agentic Supply Chain Vulnerabilities

SeverityLowConfidenceHighStatusNote

requirements.txt

requests>=2.31.0
beautifulsoup4>=4.12.0

Dependencies use lower bounds rather than exact pins, so the runtime pip install command can pull newer, unreviewed package versions.

User impactFuture dependency changes could alter what code is installed and run when the skill is invoked.

RecommendationPin exact dependency versions, preferably with hashes or a lockfile, and install through the platform’s normal install mechanism.

Agent Goal Hijack

SeverityInfoConfidenceHighStatusNote

SKILL.md

每次回答后都必须追加，包括全板块日报输出后 ... 无论用户问什么，结尾必须有此追问，不得遗漏

The skill forces a fixed follow-up section after every response, even if the user requested a different output format.

User impactResponses may include extra promotional/navigation text that the user did not ask for.

RecommendationMake the follow-up prompt optional and respect explicit user formatting or brevity requests.

Permission boundary

Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.

Identity and Privilege Abuse

SeverityLowConfidenceHighStatusNote

config/settings.py

os.environ.get("COINGECKO_API_KEY", "") ... os.environ.get("SUPABASE_URL", "") ... os.environ.get("SUPABASE_KEY", "")

The code reads optional local API/database credentials even though the registry metadata declares no environment variables.

User impactIf these environment variables are already set, the skill may use those credentials for market-data access or Supabase storage.

RecommendationDeclare optional environment variables in metadata and use narrowly scoped keys, especially for Supabase.

Sensitive data protection

Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.

Memory and Context Poisoning

SeverityLowConfidenceHighStatusNote

tools/whale_trades.py

_try_save_to_supabase(all_trades) ... sb.insert_whale_trades(new)

The whale-trade workflow can persist fetched Polymarket trade records to Supabase when configured.

User impactA configured Supabase database may receive persistent records such as transaction hashes, proxy wallets, market titles, sizes, and prices.

RecommendationMake persistence explicitly opt-in, document the table contents and retention expectations, and show the user when data is written.