- Home
- Skills
- Prompt & Skill Engineering
- pre send eq check
More screenshots
Works with the AI Agents you already use
pre send eq check
Verify draft accuracy and tone against approved facts with a rigorous PASS/FAIL rubric and fact-checked rewrites.
$19
pre send eq check
Example session with this skill installed
Check this draft: 'Join us on Jan 10 for our $99 workshop at bit.ly/webinar.' Approved facts: F1: Workshop date is Jan 12. F2: Price is $99. F3: Link is bit.ly/webinar-12. Audience: Developers.
- Read your context and instructions
- Compiled the pre send eq
Pre-Send EQ Check — NOT a live eqbuilder.dev score.
Original verdict: FAIL (Factual Contradiction)
Fact-check table
| Draft Claim | Category | Fact ID | Excerpt | Status | Recommended Action |
|---|---|---|---|---|---|
| Jan 10 | Date | F1 | Jan 12 | CONTRADICTED | Change to Jan 12 |
| $99 | Price | F2 | $99 | SUPPORTED | Preserve |
| bit.ly/webinar | URL | F3 | bit.ly/webinar-12 | CONTRADICTED | Update to correct slug |
Human-likeness rubric
- Tone Consistency: 2/2 - Professional and direct.
- Clarity: 2/2 - Clear value proposition.
- Empathy: 1/2 - Functional but slightly brief.
- Conciseness: 2/2 - No fluff.
- Engagement: 1/2 - Standard call to action.
- Total: 8/10. Rubric Verdict: PASS (Override by Factual FAIL).
Rewrite
Join us on Jan 12 for our $99 workshop at bit.ly/webinar-12.
Rewrite verification
- Factual findings: All claims (Date: Jan 12, Price: $99, URL: bit.ly/webinar-12) now map to F1, F2, and F3.
- Rubric Score: 8/10.
- Overall Verdict: PASS.
Changes and questions
- Removed: Incorrect date (Jan 10) and truncated URL.
- Preserved: Price and intent.
- Notes: Verified link destination against F3.
Correction ledger: Not recorded.
Checked only against the buyer-supplied approved facts. This does not independently verify those facts or guarantee recipient response. Review the final draft before sending.
Next steps
- Copy the rewrite for use in your outbound tool.
- Confirm the bit.ly link redirects to the intended landing page.
- Update your source facts if the Jan 10 date was actually the intended one.
Connects securely to your tools. The creator never sees your data.
About this skill
Drafting messages that include hallucinated prices, broken URLs, or tone-deaf phrasing can damage professional relationships. This skill acts as a rigorous gatekeeper, validating your drafts against a set of approved facts before you hit send.
What it does
- Fact verification maps every claim in your draft to a list of buyer-supplied approved facts.
- Inventory tracking extracts and inspects every price, date, duration, and URL for accuracy.
- Rubric scoring assesses human-likeness across five dimensions to ensure the tone is appropriate.
- Fact-preserving rewrites remove unsupported claims while keeping the original intent and approved data.
- Detailed auditing provides a status for every claim: Supported, Unsupported, Contradicted, or Ambiguous.
How it works
- Input phase where you provide the draft, the approved fact list, and the intended audience or tone.
- Analysis phase where the agent splits the draft into claims and inventories all numerical and link data.
- Evaluation phase applying a five-dimension rubric and a factual override check to issue a verdict.
- Correction phase delivering a rewritten draft and a verification report to confirm the new version passes.
Frameworks & tools
This skill is a standalone logic engine that requires no external APIs or network calls. It is designed for use within LLM environments like Claude Code or Cursor using local Markdown reference files.
Why this beats prompting it yourself
Generic prompts often "hallucinate" missing information to make a sentence sound better. This skill uses a strict "Lack of evidence means unsupported" rule and an explicit mapping system that prevents the AI from inventing replacement facts or guessing at ambiguous dates.
Use cases
- Reviewing sales outreach to ensure pricing and promo URLs match the official sheet.
- Validating internal status reports against project milestone data and deadlines.
- Checking customer support templates for factual accuracy and human-like tone.
- Auditing event invitations to verify dates, locations, and registration links.
Known limitations
Does not perform external research or live URL testing. Requires the user to provide the source of truth facts manually. Use for one draft at a time.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Prompt & Skill Engineering.
Cannot perform external research, browse links, or call live scoring APIs. Requires user-supplied facts to function. Process one draft at a time.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 4 days ago
- Accounts
Gmail
Permissions
Allowed Hosts