Agent skill
purple-safety-check
Pre-send and pre-copy safety review for outbound claims, spam triggers, and compliance risk.
Filed under Outbound email.
From rasulshaikh/outbound-skill-pack · 7 skill entries · 0 · pushed 2026-09-29
What it does when it runs
Pre-send and pre-copy safety review for outbound claims, spam triggers, and compliance risk. Use before humanizing emails or exporting to a sequencer.
Automated analysis of the skill and the 0 files bundled beside it. A skill’s own description is written to be selected by an agent, so it describes the job and not the dependencies.
- Keys and connectors you must supply
- None found.
- Hosts it reaches
- No third-party host appears in the skill or its bundled files.
- Tool permissions it declares
- No
allowed-toolsin the frontmatter. It only issues instructions, so there is nothing to bound. - Actions present in the files
- None. Instructions only.
Install it
View source on GitHub ↗git clone --depth 1 --filter=blob:none --sparse https://github.com/rasulshaikh/outbound-skill-pack.git /tmp/outbound-skill-pack git -C /tmp/outbound-skill-pack sparse-checkout set "skills/purple-safety-check" mkdir -p ~/.claude/skills/purple-safety-check cp -R "/tmp/outbound-skill-pack/skills/purple-safety-check/." ~/.claude/skills/purple-safety-check/
Picked up without a restart. A project skill of the same name is shadowed by your personal one. For one repository only, swap ~/.claude/skills for .claude/skills. Claude Code docs ↗
The folder is the same in every client that implements the format — 46 of them — so if yours is not above, only the destination changes.
The skill
Source on GitHub ↗Reproduced in full from rasulshaikh/outbound-skill-pack/blob/9516d4bfa03b15648816760032ec8035370ca9f4/skills/purple-safety-check/SKILL.md, which is licensed MIT (repository). 238 words, 9 headings.
Purple Safety Check
"Purple" here means cautious send hygiene: not illegal advice, just practical risk reduction for cold outbound.
When to use
- Before locking subject lines and bodies
- After any claim about savings, results, or customer names
- Before sequencer export QA
Checklist
Run every item. Mark PASS / FLAG / FAIL.
- Claim honesty: Every numeric or named claim has a cited source or is removed.
- Impersonation: No pretending to be a mutual contact, investor, or teammate.
- Sensitive data: No private emails scraped from breaches; no personal health or finance details.
- Spam trigger scan: Excessive caps, fake urgency, misleading "Re:" / "Fwd:", money-guarantee language.
- Unsubscribe / identity: From-name matches real sender; landing pages and physical address ready if required by the channel.
- Suppression: Known customers, competitors, and do-not-contact domains listed for exclude.
- Jurisdiction caution: Flag regions or industries that need extra legal review (finance, healthcare, kids). Do not give legal advice; escalate.
Output format
## Purple safety report
| Item | Result | Note |
| --- | --- | --- |
| Claim honesty | PASS/FLAG/FAIL | ... |
## Blockers (FAIL)
- ...
## Soft flags
- ...
## Cleared to proceed?
yes / no (list required fixes)
Guardrails
- FAIL blocks export. FLAG needs a human decision.
- This skill does not replace counsel. It is an operator checklist.
Other skills for the same job
Different authors, same problem. Matched on the words in the skill name, across every library in the catalogue except this one.
- keyword-cannibalization-check by irinabuht12-oss · 3,698
- claim-check by pmalliance · 73
- message-consistency-check by pmalliance · 73
- evaluation-pipeline-check by zime-ai · 17
- negotiation-pipeline-check by zime-ai · 17
- poc-pilot-pipeline-check by zime-ai · 17
- prospect-pipeline-check by zime-ai · 17
- qualify-pipeline-check by zime-ai · 17
Need help setting it up?
This page tells you what purple-safety-check does and what it needs. Cheetah builds the agent setup it runs inside: data, CRM, sequencing and the guardrails.
Book a call →The directory stays free. There is nothing gated behind this.