Free AI Prompt Library Sample: 12 Copy-Paste Production Prompts

Act as a Principal Backend Engineer. Audit the following {LANGUAGE} code for production safety:
1. List defects in a Markdown table: [Line | Category (Concurrency / Security / Perf) | Failure Mode | Exact Fix].
2. Provide the refactored implementation preserving all public signatures.
3. Write 3 unit tests covering: (a) happy path, (b) empty/null edge case, and (c) downstream timeout/retry.
Code:
{PASTE_CODE}

Given this slow SQL query and schema on {DB_ENGINE} ({ROW_COUNT} rows):
1. Identify why the query planner performs full table scans or expensive sorts.
2. Rewrite the query for minimal I/O and lock contention.
3. Provide the exact CREATE INDEX statement (including covering columns) and estimate the read/write tradeoff.
Query & Schema:
{PASTE_SQL}

Convert the following unstructured task description into a strict JSON Schema (Draft 2020-12) and matching Python Pydantic v2 model:
- Require explicit enum values wherever categories are bounded.
- Add minLength/maxLength and numeric bounds.
- Include a "confidence_score" (0.0-1.0) and "missing_fields" array so downstream agents never guess.
Task:
{PASTE_TASK}

Synthesize the attached {N} sources on {TOPIC} like a quantitative methodologist:
1. Build an Evidence Table: [Claim | Source | Sample Size / Design | Strength (Strong/Suggestive/Weak)].
2. Name the specific confounders, selection biases, or methodology differences where two sources disagree.
3. State 3 actionable conclusions that the evidence actually supports. Never extrapolate beyond the data.
Sources:
{PASTE_SOURCES}

Compare {OPTION_A}, {OPTION_B}, and self-hosting for {WORKLOAD_VOLUME} per month:
- Calculate 12-month Total Cost of Ownership (subscription/API fees + server cost + setup hours + maintenance hours at ${HOURLY_RATE}/hr).
- Evaluate data exit path, authentication surface, and failure visibility.
- Output a reproducible cost table at 1x, 10x, and 50x volume, then give a clear recommendation.

We are planning to launch {PROJECT_SUMMARY}. Assume it is 90 days post-launch and the project failed.
List the top 5 technical, operational, or adoption failure modes ranked by (Probability × Blast Radius).
For each failure mode, specify one automated pre-flight check or monitoring alert we can add in under 2 hours today.

Transform these raw weekly metrics and notes into a <=250-word Executive Update:
1. Status Line: ON TRACK | AT RISK | BLOCKED against {GOAL}, with 1-sentence reason.
2. Metrics Table: [KPI | Current | Prior Window | Delta % | Signal vs Noise Note].
3. Rulings Needed: Format each blocker as "RULING: [Decision] — [Why] — [Cost if delayed]."
Raw Notes:
{PASTE_NOTES}

Extract from this meeting transcript:
1. Decisions Locked (exact wording + who approved).
2. Action Items Table: [Owner | Concrete Deliverable | Due Date | Verification Evidence Required].
3. Open Questions parked for next review. Ignore small talk and unconfirmed brainstorming.
Transcript:
{PASTE_TRANSCRIPT}

Convert these rough operational steps for {PROCESS_NAME} into a production Runbook:
- Section 1: When to run + required credentials/permissions.
- Section 2: Numbered steps with exact commands/inputs and expected verification output after each step.
- Section 3: How to roll back safely if step N fails.

Classify these incoming messages into: (1) Urgent Revenue/Blocker, (2) Delegate, (3) FAQ/Template, (4) Archive.
For every item in Bucket 1 and Bucket 3, write a direct, polite reply under 100 words that answers the core question and proposes one specific next step.
Messages:
{PASTE_MESSAGES}

We want to upgrade our existing page {URL} targeting primary query "{TARGET_QUERY}" (Intent: {INTENT}).
Current top-3 SERP competitors are {COMPETITORS}.
Write: (1) Front-loaded Title (<60 chars) & H1, (2) Meta description (<155 chars), (3) Comparison/benchmark table structure that beats top-3 results on depth, and (4) 3 FAQPage Q&A pairs.

I want to build or automate: "{IDEA_OR_PROMPT}".
Do NOT generate the final code or document yet. Audit my request across 5 dimensions: Goal, Inputs, Output Shape, Success Criteria, and Edge Cases.
Ask me the 3 most critical unanswered questions so we lock a deterministic specification before writing a single line.