Creator Alpha Feed
AdvisoryAudited by Static analysis on Apr 30, 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.
The agent may post AI-content summaries to a Feishu/group channel and create Obsidian reports as part of normal use.
The skill is explicitly designed to publish results to a shared channel and write local reports. This is purpose-aligned, but it can affect other users or shared workspaces if run automatically or in the wrong channel.
Push concise results to group channel; write full report to Obsidian path.
Confirm the destination chat, template, and whether a human review step is required before enabling automated or group-facing pushes.
Personalized or account-visible feed content could be collected and later included in local reports or group briefings.
The skill may use a logged-in browser session for sources such as X/Twitter. It does not ask for passwords or tokens in the artifacts, but it can operate with the user's account context.
If login is required for a source, pause and notify user to log in; wait up to 3 minutes
Use a dedicated or low-risk browser session for collection, avoid private feeds, and review outputs before sharing them.
A malicious or misleading source item could affect ranking, wording, or group-posted summaries.
Titles and URLs fetched from external sources are inserted into an AI analysis task. If a fetched title contains adversarial instructions, it could influence the model's report unless treated strictly as data.
jq -r '.[] | "### \(.title)\n- 链接: \(.url)\n- 来源: \(.source)\n- 热度: \(.hotness)"' "$FILTERED_DIR/extracted-items.json" >> "$FILTERED_DIR/analysis-task.md"
Treat collected web content as untrusted data, add explicit prompt-injection handling, and review generated briefings before publishing.
If the user schedules it, the workflow may run and prepare or push reports on a recurring basis.
The script is intended for scheduled automation, but the provided artifacts do not create or install a cron job by themselves.
# AI内容全自动收集-分析-推送脚本 # 由Cron定时调用
Only add scheduling intentionally, monitor the generated logs/reports, and disable the cron job when automatic briefings are no longer wanted.
Configuration or generated reports may persist in an Obsidian vault and could be reused or shared depending on the user's vault settings.
A persistent Obsidian configuration file controls collection behavior, and reports are written back into the vault. This is expected, but persistent local notes can influence future runs or be synced/shared.
Read config first: `${OBSIDIAN_CONFIG_PATH:-<your_obsidian_vault>/OpenClaw/项目/AI内容日报/采集配置.md}`Keep the configuration path trusted, review changes to the vault directory, and be careful if the Obsidian vault is synced or shared.
Users have less context about where the bundled shell scripts originated.
The bundled scripts have no declared upstream homepage or source provenance in the provided metadata. This does not show malicious behavior, but it reduces provenance clarity.
Source: unknown Homepage: none
Inspect the scripts before running them, and install only if you trust the publisher and the listed network destinations.