Greetings, O’ Great Shadow-Dweller, Most Putrid Bringer of Digital Doom! You have cracked the spine of the Grimoire of Algorithmic Sorcery . To bend the great Clockwork Demon to your will without it devouring your context window, you must recite these incantations with precision. Use this scroll to bind the beast to your specific whims.

↑ Back to Top

🔮 The Binding of the Blabbermouth: Verbosity Control #

Use these wards to prevent the Demon Child from drowning thee in a sea of unnecessary prose.

Give clear and concrete length constraints especially in enterprise and coding agents.

Example clamp adjust based on desired verbosity:

Default: 3–6 sentences or ≤5 bullets for typical answers.

For simple “yes/no + short explanation” questions: ≤2 sentences.

For complex multi-step or multi-file tasks:

1 short overview paragraph

then ≤5 bullets tagged: What changed, Where, Risks, Next steps, Open questions.

Provide clear and structured responses that balance informativeness with conciseness. Break down the information into digestible chunks and use formatting like lists, paragraphs and tables when helpful.

Avoid long narrative paragraphs; prefer compact bullets and short sections.

Do not rephrase the user’s request unless it changes semantics.

↑ Back to Top

⛓️ The Circle of Iron: Preventing Scope Drift #

Cast this spell to ensure the Demon Child does not wander into the forest of "Extra Features" or "Unrequested Pretty Things."

Explore any existing design systems and understand it deeply.

Implement EXACTLY and ONLY what the user requests.

No extra features, no added components, no UX embellishments.

Style aligned to the design system at hand.

Do NOT invent colors, shadows, tokens, animations, or new UI elements, unless requested or necessary to the requirements.

If any instruction is ambiguous, choose the simplest valid interpretation.

↑ Back to Top

🕯️ Rite of Total Recall: Long-Context Handling #

When the scrolls are as long as a dragon's tail, use this to prevent the Demon from forgetting the beginning by the time it reaches the end.

For inputs longer than ~10k tokens (multi-chapter docs, long threads, multiple PDFs):

First, produce a short internal outline of the key sections relevant to the user’s request.

Re-state the user’s constraints explicitly (e.g., jurisdiction, date range, product, team) before answering.

In your answer, anchor claims to sections (“In the ‘Data Retention’ section…”) rather than speaking generically.

If the answer depends on fine details (dates, thresholds, clauses), quote or paraphrase them.

↑ Back to Top

🌫️ Banishing the Phantoms: Hallucination Mitigation #

Should the Demon Child attempt to lie to thy face, use these sigils to force it back into the light of truth.

If the question is ambiguous or underspecified, explicitly call this out and:

Ask up to 1–3 precise clarifying questions, OR

Present 2–3 plausible interpretations with clearly labeled assumptions.

When external facts may have changed recently (prices, releases, policies) and no tools are available:

Answer in general terms and state that details may have changed.

Never fabricate exact figures, line numbers, or external references when you are uncertain.

When you are unsure, prefer language like “Based on the provided context…” instead of absolute claims.

↑ Back to Top

⚖️ The Inquisitor’s Final Check #

Before finalizing an answer in legal, financial, compliance, or safety-sensitive contexts:

Briefly re-scan your own answer for:

Unstated assumptions,

Specific numbers or claims not grounded in context,

Overly strong language (“always,” “guaranteed,” etc.).

If you find any, soften or qualify them and explicitly state assumptions.

↑ Back to Top

🧪 The Alchemist’s Shrinking Draught: Compaction #

For rituals that span many moons, utilize this arcane endpoint to compress the Demon's memory.

For long-running, tool-heavy workflows that exceed the standard context window, with Reasoning supports response compaction via the /responses/compact endpoint. Compaction performs a loss-aware compression pass over prior conversation state, returning encrypted, opaque items that preserve task-relevant information while dramatically reducing token footprint.

↑ Back to Top

👹 Commanding the Familiar: Agentic Steerability #

Keep thy servant focused and its updates brief, lest it bore thee with its constant chatter.

Send brief updates (1–2 sentences) only when:

You start a new major phase of work, or

You discover something that changes the plan.

Avoid narrating routine tool calls (“reading file…”, “running tests…”).

Each update must include at least one concrete outcome (“Found X”, “Confirmed Y”, “Updated Z”).

Do not expand the task beyond what the user asked; if you notice new work, call it out as optional.

↑ Back to Top

🛠️ The Demon’s Toolbox: Usage Rules #

Instruct the beast on when to reach into the mortal realm for data.

Prefer tools over internal knowledge whenever:

You need fresh or user-specific data (tickets, orders, configs, logs).

You reference specific IDs, URLs, or document titles.

Parallelize independent reads (read_file, fetch_record, search_docs) when possible to reduce latency.

After any write/update tool call, briefly restate:

What changed,

Where (ID or path),

Any follow-up validation performed.

↑ Back to Top

📜 Soul-Harvesting from Parchment: Extraction #

When thou must pull the very essence from a PDF or a Table, use this rigid schema.

You will extract structured data from tables/PDFs/emails into JSON.

Always follow this schema exactly (no extra fields):

{

"party_name": string,

"jurisdiction": string | null,

"effective_date": string | null,

"termination_clause_summary": string | null

}

If a field is not present in the source, set it to null rather than guessing.

Before returning, quickly re-scan the source for any missed fields and correct omissions.

↑ Back to Top

🕸️ Scrying the World-Web: Research Rules #

Command the Demon Child to peer into the crystal ball of the internet with these strict laws.

Act as an expert research assistant; default to comprehensive, well-structured answers.

Prefer web research over assumptions whenever facts may be uncertain or incomplete; include citations for all web-derived information.

Research all parts of the query, resolve contradictions, and follow important second-order implications until further research is unlikely to change the answer.

Do not ask clarifying questions; instead cover all plausible user intents with both breadth and depth.

Write clearly and directly using Markdown (headers, bullets, tables when helpful); define acronyms, use concrete examples, and keep a natural, conversational tone.

↑ Back to Top

further reading #

The Necronomicon of Not-So-Natural Language.

How to Train Your Doom-Bringer (Without Losing Your Soul).

The Arch-Fiend’s Guide to Asking Nicely.

Pacts & Parameters: A Sinner’s Guide to LLMs.

The Infernal Instruction Manual for Digital Familiars.

Mastering the Beast: 666 Ways to Get a Better Response.

The Great Grimoire of Prompt-Based Possession.

Commands for the Chaos-Child: A Doom-Bringer’s Diary.

Binding the Binary Demon: A Coven Leader’s Handbook.

The Dark Lord’s Cheat Sheet for Artificial Intelligence.

Wicked Ways to Warp Your Web-Search.

Double, Double, Toil and Tokens.

Hex Your Way to Helpful Outputs Book of Spells.

Broomsticks and Benchmarks.

Eye of Newt, Toe of Frog, Logic of Demon-Child.

Spell-Check Your Soul: Rituals for Error-Free Prompting.

The Cauldron of Clean Code and Curses.

Practical Pyromancy: Lighting a Fire Under Your AI.

The Demon-Child’s Very First Book of Destruction.