feapder

PassAudited by ClawScan on May 10, 2026.

Overview

This is a coherent, instruction-only feapder development aid; the main cautions are user-run crawler/database workflows and unsafe eval patterns in vendored reference code, not hidden automatic behavior.

This skill appears safe to install as a feapder reference assistant. Before running any generated spider or CLI command, confirm the target website, Redis key, database name, and credentials; avoid pasting real cookies or secrets into shared prompts or committed files; and be cautious with feapder 1.9.2 internals that use eval on task/debug strings.

Findings (3)

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

If a project runs this kind of code against attacker-controlled Redis/MySQL task data, a crafted task string could execute Python code.

Why it was flagged

The vendored feapder reference implementation contains dynamic evaluation of task data. It is not auto-run by this instruction-only skill, but it is an unsafe pattern if the upstream code or copied logic processes untrusted task strings.

Skill content
tasks = [eval(task) for task in tasks]
Recommendation

Do not copy eval-based parsing into new spider code. Prefer json.loads or ast.literal_eval, and only run feapder 1.9.2 task/debug flows against trusted data or patched code.

What this means

Running generated batch crawler code against the wrong database or Redis namespace could mark many tasks complete or failed incorrectly.

Why it was flagged

The documented BatchSpider workflow includes updating task state in backing storage. This is expected for feapder batch crawlers, but it is still a database/queue mutation when generated code is run.

Skill content
yield self.update_task_batch(request.task_id, 1) # 更新任务状态为1
Recommendation

Review redis_key, task_table, task_state, and database settings before running generated spiders, and test against a non-production environment first.

What this means

Copying authenticated browser curl commands into prompts, scripts, or shell history can expose session cookies.

Why it was flagged

The debugging documentation shows raw curl commands that may include session cookies. This is a normal crawler-debugging technique, but cookies are sensitive credentials.

Skill content
feapder shell --curl 'https://www.baidu.com/' ... -H 'Cookie: PSTM=...; BDUSS=...'
Recommendation

Redact cookies and authorization headers before sharing prompts or committing code, and use temporary test sessions when debugging authenticated pages.