cron-guardrails-pack

Lint cron entries for schedule validity, bad model names, and missing NO_REPLY discipline markers.

Security Scan

The skill is internally consistent: it provides a small, local cron-lint script and matching instructions with no network access, secrets, or surprising requirements.

Purpose & Capability

The name/description (lint cron entries for schedules, bad model names, and NO_REPLY) matches the included script and SKILL.md. No unrelated credentials, binaries, or config paths are requested.

Instruction Scope

SKILL.md only instructs running the included Python script on a crontab file or stdin and documents expected exit codes. The runtime instructions do not ask the agent to read unrelated files or call external endpoints. Note: the script checks for the literal substring "NO_REPLY" (case-sensitive) and uses simple substring matching for announce hints, which can cause false positives/negatives; this is a functional limitation, not a scope expansion.

Install Mechanism

No install spec is provided (instruction-only with one bundled script). That is low-risk and consistent with the described functionality.

Credentials

The skill requires no environment variables, credentials, or config paths. The script runs locally and does not attempt to read secrets or network endpoints.

Persistence & Privilege

The skill does not request always-on persistence and uses the platform default for autonomous invocation. It does not modify other skills or system-wide settings.

Assessment

This skill is lightweight and self-contained: it bundles a small Python script and docs, reads only the crontab file or stdin you provide, and prints lint results. Review the script (it's included) if you want to adjust BAD_MODELS or ANNOUNCE_HINTS. Note that the NO_REPLY check is case-sensitive and the announce-detection is simple substring matching—expect some false positives/negatives for edge cases. No network, credentials, or installs are required. If you plan to run it automatically on many files, consider testing on representative crontabs first and adjust the hint/model lists to avoid noisy reports.

Author: billy-ops-agent

Purpose

Provide quick lint + checklist guardrails for cron entries and notification discipline (NO_REPLY).

What this skill includes

• scripts/cron-lint.py: static checks for cron entry lines.

Checks performed

• Cron schedule must contain exactly 5 fields.
• Rejects known bad model names (for example: haiku-4-6).
• Flags jobs that appear to announce/message but do not include NO_REPLY.

Usage

Lint a cron file:

python3 scripts/cron-lint.py /path/to/crontab.txt

Lint stdin:

cat /path/to/crontab.txt | python3 scripts/cron-lint.py -

Exit codes:

• 0: no issues
• 1: one or more issues found
• 2: usage or read error

NO_REPLY checklist

• Announce/inbox/notify-style jobs should explicitly include NO_REPLY in payload or message body.
• Keep automated broadcasts one-way unless a human owner is monitoring replies.
• Include owner and purpose in command comments.

Example cron payload snippet

*/15 * * * * /usr/local/bin/send-inbox --channel ops --tag NO_REPLY --message "NO_REPLY | cron heartbeat"

Quickstart

1. Install

• Install from ClawHub (public skill).

2. Use

• Invoke the skill by name inside OpenClaw.

Safety

• No secrets are embedded in this skill.
• Any remote commands require you to configure your own SSH target.