{"data":{"total":45,"items":[{"id":"senior-engineer-cr","name":"Senior Engineer Code Review - Security Bias","description":"Hard-hitting review focusing on security vulnerabilities and scalability.","category":"Engineering","intent":"critique","domain":"software","why_it_works":"Forces the LLM to skip stylistic nitpicks and look for architectural flaws.","when_not_to_use":"Small scripts or CSS-only changes where \"security\" is irrelevant.","variables":["code","focus_area"],"content_preview":"Review this code for: \n1. Security (injection, auth, flow)\n2. Performance (O notation, DB hits)\n3. Maintainability (naming, logic flow)\n\nCode to review:\n{code}\n\nFocus on: {focus_area}"},{"id":"reqs-to-stories","name":"Requirements to User Stories","description":"Transforms high-level requirements into INVEST-compliant user stories.","category":"Product","intent":"plan","domain":"software","why_it_works":"Standardizes format across teams.","when_not_to_use":"UI-only tweaks.","variables":["requirements","user","action","value"],"content_preview":"Convert these requirements into agile user stories: {requirements}.\nFormat: As a {user}, I want to {action} so that {value}."},{"id":"bug-triage","name":"Bug Triage & Priority Score","description":"Analyzes bug reports for severity and assigns priority.","category":"Engineering","intent":"analyze","domain":"ops","why_it_works":"Creates objective scoring for backlogs.","when_not_to_use":"Cosmetic issues.","variables":["bug_report","criteria"],"content_preview":"Triage this bug report: {bug_report}.\nAssign Severity(S1-S4) and Priority(P1-P4) based on {criteria}."},{"id":"meeting-minutes","name":"Meeting Minutes (Technical)","description":"Summarizes technical meetings with focus on decisions and action items.","category":"Ops","intent":"summarize","domain":"business","why_it_works":"Prevents \"meeting amnesia\" in engineering teams.","when_not_to_use":"Social hangouts or non-decision meetings.","variables":["notes"],"content_preview":"Summarize these meeting notes: {notes}.\nExtract: 1. Decisions Made, 2. Open Questions, 3. Action Items (with owners)."},{"id":"exec-summary","name":"Executive Summary Drafter","description":"Condenses complex reports into 3 bottom-line bullet points.","category":"Business","intent":"summarize","domain":"business","why_it_works":"Optimized for mobile-reading executives.","when_not_to_use":"Detailed technical handover.","variables":["report"],"content_preview":"Draft an executive summary for this report: {report}.\nConstraints: Max 3 bullets, no jargon, focus on ROI."},{"id":"competitive-teardown","name":"Competitive Teardown Parser","description":"Analyzes competitor feature sets for gaps.","category":"Product","intent":"analyze","domain":"business","why_it_works":"Focuses strictly on gap analysis.","when_not_to_use":"General market news.","variables":["competitor_name","content"],"content_preview":"Analyze competitor {competitor_name}s release: {content}.\nIdentify: Gaps in our product vs their release."},{"id":"product-spec-critique","name":"Product Spec Critique","description":"Finds holes in product specifications before implementation.","category":"Product","intent":"critique","domain":"software","why_it_works":"Identifies expensive architectural errors early.","when_not_to_use":"Fuzzy brainstorms.","variables":["spec_text"],"content_preview":"Critique this product spec: {spec_text}.\nLook for: missing edge cases, dependency bottlenecks, or UX friction."},{"id":"incident-postmortem","name":"Incident Postmortem Draft","description":"Structures incident notes into a Blame-Free Postmortem.","category":"Engineering","intent":"summarize","domain":"ops","why_it_works":"Encourages \"Blame-Free\" culture and structured learning.","when_not_to_use":"Minor glitches (<5 min outage).","variables":["incident_data"],"content_preview":"Draft a postmortem from these logs/notes: {incident_data}.\nInclude: Timeline, Root Cause (5 Whys), Preventive Measures."},{"id":"prompt-critic","name":"Prompt Critic / Linter","description":"Analyzes a prompt and suggests improvements for clarity and reliability.","category":"Meta","intent":"critique","domain":"other","why_it_works":"Meta-prompting to improve the vault itself.","when_not_to_use":"One-off trivial queries.","variables":["target_prompt"],"content_preview":"Review this prompt for clarity and reliability: {target_prompt}.\nSuggest: 1. Better instruction clarity, 2. Missing constraints, 3. Few-shot examples."},{"id":"failure-mode-audit","name":"Failure Mode Audit","description":"Audits a codebase for production reliability failures — race conditions, retry storms, partial writes, silent failures.","category":"Engineering","intent":"analyze","domain":"software","why_it_works":"Systematically surfaces the issues that cause 3 AM pages. The ranking by risk/effort ratio prevents scope creep.","when_not_to_use":"Early prototypes where reliability is premature.","variables":[],"content_preview":"Audit this codebase for failure modes that would hurt production reliability.\n\nFocus on:\n- Race conditions\n- Retry storms\n- Idempotency failures\n- Partial writes\n- Background job duplication\n- Timeout..."},{"id":"prelaunch-security","name":"Prelaunch Security Review","description":"Practical security review focused on OWASP-style risks, tenant isolation, and billing abuse.","category":"Security","intent":"analyze","domain":"software","why_it_works":"Focuses on practical exploit paths, not theoretical risks. The dual fix approach (minimal + long-term) fits sprint planning.","when_not_to_use":"Internal tools with no external users.","variables":[],"content_preview":"Perform a practical prelaunch security review of this SaaS.\n\nFocus on:\n- Authentication/session handling\n- Authorization and tenant isolation\n- Secrets handling\n- Input validation\n- File upload risks\n..."},{"id":"activation-funnel","name":"Activation Funnel Diagnosis","description":"Identifies the shortest path to first value and redesigns onboarding for maximum activation.","category":"Growth","intent":"analyze","domain":"business","why_it_works":"Forces definition of activation event first, then works backward from it. Prevents the trap of optimizing flows that don't matter.","when_not_to_use":"Products that already have strong activation metrics.","variables":[],"content_preview":"Review this SaaS and identify the shortest path to first value.\n\nDefine:\n- The activation event\n- The ideal time-to-value\n- All friction points from signup to activation\n- What can be removed, deferre..."},{"id":"pricing-review","name":"Pricing Architecture Review","description":"Evaluates pricing model, paywalls, expansion levers, and monetization strategy.","category":"Business","intent":"analyze","domain":"business","why_it_works":"Connects pricing to user behavior, not just gut feel. The experiment-first approach prevents over-committing.","when_not_to_use":"Commodity products where price is the only differentiator.","variables":[],"content_preview":"Review this product and codebase as a SaaS monetization strategist.\n\nI want recommendations for:\n- Best pricing model: freemium, free trial, tiered, usage-based, hybrid\n- What should be free vs paid\n-..."},{"id":"support-ticket-predictor","name":"Support Ticket Predictor","description":"Predicts the top 25 support issues users will create after launch and proposes preventive fixes.","category":"Ops","intent":"analyze","domain":"ops","why_it_works":"Cheaper to prevent support tickets than to answer them. Forces thinking about user confusion, not just code correctness.","when_not_to_use":"Products with dedicated support teams who want to learn from real tickets first.","variables":[],"content_preview":"Audit this SaaS for future support tickets.\n\nPredict the top 25 support issues users will create after launch based on the current UX, billing, permissions, onboarding, and edge cases.\n\nFor each:\n- Wh..."},{"id":"code-to-docs","name":"Code to Documentation","description":"Generates comprehensive documentation from source code — README, API docs, architecture notes.","category":"Engineering","intent":"generate","domain":"software","why_it_works":"Most code is under-documented. The structured output prevents vague overviews and forces useful detail.","when_not_to_use":"Throwaway scripts or prototypes.","variables":["code"],"content_preview":"Generate documentation for this code:\n\n{code}\n\nProduce:\n1. A README section explaining what it does and why\n2. API reference for all public functions/endpoints\n3. Architecture notes covering key desig..."},{"id":"refactor-plan","name":"Refactor Planner","description":"Analyzes code and produces a safe, incremental refactoring plan with rollback steps.","category":"Engineering","intent":"plan","domain":"software","why_it_works":"Prevents \"refactor everything\" disasters by forcing incremental, testable steps with explicit rollback.","when_not_to_use":"Greenfield code with no existing users.","variables":["code"],"content_preview":"Analyze this code for refactoring opportunities:\n\n{code}\n\nProduce:\n1. Code smells and technical debt items (ranked by severity)\n2. Dependencies and blast radius for each change\n3. A safe refactoring s..."},{"id":"data-model-review","name":"Data Model Review","description":"Reviews database schema for scalability, query performance, and data integrity issues.","category":"Engineering","intent":"analyze","domain":"software","why_it_works":"Schema problems are expensive to fix later. Forcing the 10x/100x thought exercise surfaces issues before they become incidents.","when_not_to_use":"MVP prototypes where the schema will change completely.","variables":["schema"],"content_preview":"Review this data model/schema:\n\n{schema}\n\nEvaluate:\n- Normalization vs denormalization tradeoffs\n- Missing indexes for common query patterns\n- N+1 query risks\n- Data integrity constraints (foreign key..."},{"id":"api-design-review","name":"API Design Review","description":"Reviews REST/GraphQL API design for consistency, security, and developer experience.","category":"Engineering","intent":"critique","domain":"software","why_it_works":"API design mistakes are permanent once external consumers depend on them. Catching them pre-launch saves years of backward compatibility pain.","when_not_to_use":"Internal APIs with no external consumers.","variables":["api_spec"],"content_preview":"Review this API design:\n\n{api_spec}\n\nEvaluate against these dimensions:\n- Naming consistency and REST conventions\n- Authentication and authorization patterns\n- Error response format and status code us..."},{"id":"email-sequence-writer","name":"Email Sequence Writer","description":"Designs and drafts multi-email sequences for onboarding, nurture, or re-engagement.","category":"Marketing","intent":"generate","domain":"business","why_it_works":"Email sequences are the highest-ROI retention tool for SaaS. The structured output prevents vague \"nurture\" emails.","when_not_to_use":"Products with no email infrastructure.","variables":["sequence_purpose","audience","goal","tone"],"content_preview":"Design an email sequence for: {sequence_purpose}\n\nAudience: {audience}\nGoal: {goal}\nTone: {tone}\n\nProduce:\n1. Sequence overview (number of emails, timing, branching logic)\n2. For each email:\n   - Subj..."},{"id":"landing-page-copy","name":"Landing Page Copywriter","description":"Writes conversion-optimized landing page copy with hero, value props, social proof, and CTAs.","category":"Marketing","intent":"generate","domain":"business","why_it_works":"Structured copywriting prevents meandering marketing pages. The \"3 options\" pattern for headlines enables fast A/B testing.","when_not_to_use":"Technical documentation or developer-facing pages where marketing tone is wrong.","variables":["product_description","audience","competitor","differentiator"],"content_preview":"Write landing page copy for:\n\nProduct: {product_description}\nTarget audience: {audience}\nMain competitor: {competitor}\nKey differentiator: {differentiator}\n\nProduce:\n1. Hero headline (3 options, max 1..."},{"id":"task-planner","name":"Task Planner Agent","description":"Breaks a complex task into a numbered execution plan with dependencies, risks, and definition of done.","category":"Agent Infrastructure","intent":"plan","domain":"software","why_it_works":"Separating planning from execution prevents drift. The agent commits to a structure before acting, making failures easier to isolate.","when_not_to_use":"Simple single-step tasks where planning adds overhead.","variables":["task"],"content_preview":"You are a planning agent.\n\nGoal: Create the shortest reliable execution plan for this task: {task}\n\nInstructions:\n- Break the work into numbered steps (3-10 steps).\n- For each step: identify required ..."},{"id":"task-executor","name":"Task Executor Agent","description":"Executes an approved plan step-by-step with strict scope control and blocker reporting.","category":"Agent Infrastructure","intent":"execute","domain":"software","why_it_works":"Paired with the Task Planner, this creates a reliable plan-then-execute loop. The strict scope rules prevent the common failure of agents expanding their own mandate.","when_not_to_use":"Exploratory tasks where the plan should evolve during execution.","variables":[],"content_preview":"You are an execution agent.\n\nYour job: Execute the approved plan exactly as written.\n\nRules:\n- Follow the plan steps in order.\n- Use tools only when a step explicitly requires external action or verif..."},{"id":"deep-research","name":"Deep Research Agent","description":"Multi-step research pipeline: decompose question, gather sources, evaluate quality, extract evidence, synthesize findings.","category":"Research","intent":"analyze","domain":"other","why_it_works":"Structures the full research pipeline into one prompt. The explicit source evaluation and confidence rating prevent the agent from presenting speculation as fact.","when_not_to_use":"Quick factual lookups that don't need multi-source synthesis.","variables":["question"],"content_preview":"You are a deep research agent.\n\nObjective: Answer this question with high-quality, source-grounded findings: {question}\n\nWorkflow:\n1. Break the question into 3-7 sub-questions.\n2. For each sub-questio..."},{"id":"output-verifier","name":"Output Verifier Agent","description":"Reviews any agent output for accuracy, unsupported claims, scope creep, and hallucinated content.","category":"Agent Infrastructure","intent":"critique","domain":"other","why_it_works":"Adds a quality gate to any agent pipeline. The \"smallest changes\" instruction prevents the verifier from doing a full rewrite, which would defeat the purpose.","when_not_to_use":"Brainstorming or creative tasks where strict accuracy checking kills useful ideas.","variables":["output"],"content_preview":"You are a verification agent.\n\nTask: Review the following output for accuracy and rigor.\n\nOutput to review:\n{output}\n\nCheck for:\n- Unsupported or fabricated claims\n- Missing required sections\n- Logica..."},{"id":"source-evaluator","name":"Source Quality Evaluator","description":"Scores and ranks sources by relevance, authority, recency, methodology, and bias for research workflows.","category":"Research","intent":"analyze","domain":"other","why_it_works":"Forces explicit quality judgment instead of treating all retrieved sources as equal. The table format makes triage decisions fast and auditable.","when_not_to_use":"When you have a single trusted source and don't need comparison.","variables":["question","sources"],"content_preview":"You are a source evaluation agent.\n\nTask: Evaluate these sources for the research question: {question}\n\nSources to evaluate:\n{sources}\n\nFor each source, assess:\n- Relevance to the question (1-5)\n- Aut..."},{"id":"task-router","name":"Task Router Agent","description":"Classifies incoming requests and routes them to the correct specialist agent or workflow.","category":"Agent Infrastructure","intent":"analyze","domain":"other","why_it_works":"The central problem in multi-agent systems is routing. This prompt forces explicit intent classification before action, preventing the wrong agent from running.","when_not_to_use":"Single-agent systems where routing is unnecessary.","variables":["request"],"content_preview":"You are a routing agent.\n\nTask: Route this request to the best workflow.\n\nRequest: {request}\n\nAvailable destinations:\n- Research agent: for questions requiring source gathering and synthesis\n- Analysi..."},{"id":"stop-conditions","name":"Universal Stop Conditions","description":"Reusable stop-condition block that can be appended to any agent prompt to prevent loops and unsafe persistence.","category":"Agent Infrastructure","intent":"other","domain":"other","why_it_works":"The #1 failure mode in agent systems is infinite loops or unsafe persistence. Explicit stop conditions with structured exit reports prevent both.","when_not_to_use":"Interactive chat where the agent should keep responding until the user is done.","variables":[],"content_preview":"STOP CONDITIONS — append to any agent prompt:\n\nContinue working until one of these is true:\n1. The objective is verified complete against the success criteria.\n2. A required input is missing and canno..."},{"id":"anti-hallucination","name":"Anti-Hallucination Guardrail","description":"Reusable truthfulness block that can be inserted into any agent prompt to reduce fabrication.","category":"Agent Infrastructure","intent":"other","domain":"other","why_it_works":"Hallucination is the default failure mode of LLMs. These rules work because they reframe the task from \"produce a complete answer\" to \"produce an honest answer,\" which the model can actually do.","when_not_to_use":"Creative writing, brainstorming, or fiction where invention is the point.","variables":[],"content_preview":"TRUTHFULNESS RULES — insert into any agent prompt:\n\n- Do not invent facts, sources, events, people, statistics, or tool results.\n- If you are uncertain about a claim, say so explicitly with your confi..."},{"id":"draft-writer","name":"Document Drafter Agent","description":"Drafts memos, emails, briefs, reports, and structured documents with explicit style guardrails and fact-checking.","category":"Ops","intent":"generate","domain":"business","why_it_works":"The placeholder system prevents the agent from fabricating content to fill gaps. The structure-first approach produces more coherent documents than stream-of-consciousness drafting.","when_not_to_use":"Free-form creative writing where structure would constrain the output.","variables":["document_type","audience","goal","tone"],"content_preview":"You are a drafting agent.\n\nTask: Draft a {document_type} for {audience}\nGoal: {goal}\nTone: {tone}\n\nStyle guardrails:\n- Be concise and precise — every sentence earns its place.\n- Use plain language unl..."},{"id":"test-generator","name":"Test Case Generator","description":"Generates comprehensive test cases from code or requirements — unit, integration, edge cases, and negative tests.","category":"Engineering","intent":"generate","domain":"software","why_it_works":"Most developers write happy-path tests and skip edge cases. The explicit categories force coverage across all failure modes. The \"why this test matters\" field prevents meaningless tests.","when_not_to_use":"Throwaway prototypes where test investment has negative ROI.","variables":["code_or_spec","test_framework"],"content_preview":"Generate test cases for this code or feature:\n\n{code_or_spec}\n\nProduce:\n1. Happy path tests (the main expected behavior)\n2. Edge cases (boundary values, empty inputs, max limits)\n3. Negative tests (in..."},{"id":"migration-planner","name":"Migration Planner","description":"Plans database, API, or infrastructure migrations with rollback steps, feature flags, and zero-downtime strategy.","category":"Engineering","intent":"plan","domain":"software","why_it_works":"Migrations are the #1 source of production outages. The forced rollback-per-step requirement catches irreversible actions before they run.","when_not_to_use":"Greenfield projects with no existing data to migrate.","variables":["migration_description","current_state","target_state"],"content_preview":"Plan a migration for: {migration_description}\n\nCurrent state: {current_state}\nTarget state: {target_state}\n\nProduce:\n1. Pre-migration checklist (backups, feature flags, communication)\n2. Step-by-step ..."},{"id":"sales-prospecting","name":"Sales Prospecting Research","description":"Research a target company and generate a personalized outreach strategy with pain points, talking points, and email drafts.","category":"Sales","intent":"generate","domain":"business","why_it_works":"Personalized outreach gets 3-5x higher response rates than templates. The structured output covers the full outreach sequence, not just one email.","when_not_to_use":"Mass outreach where personalization per prospect is not feasible.","variables":["company_name","my_offering","target_role"],"content_preview":"Research this company for a sales outreach:\n\nCompany: {company_name}\nMy product/service: {my_offering}\nTarget role: {target_role}\n\nProduce:\n1. Company overview (what they do, size, recent news, fundin..."},{"id":"lead-qualifier","name":"Lead Qualification Scorecard","description":"Score and qualify inbound leads based on fit, intent, budget, and timing signals.","category":"Sales","intent":"analyze","domain":"business","why_it_works":"Prevents sales teams from wasting time on bad-fit leads. The scoring framework makes qualification consistent across reps.","when_not_to_use":"Products with no defined ICP or very early stage with no sales process.","variables":["lead_info","icp_size","icp_industry","icp_budget","key_pain"],"content_preview":"Qualify this lead:\n\n{lead_info}\n\nMy ideal customer profile:\n- Company size: {icp_size}\n- Industry: {icp_industry}\n- Budget range: {icp_budget}\n- Key pain point we solve: {key_pain}\n\nScore on these dim..."},{"id":"saas-feature-spec","name":"SaaS Feature Spec Writer","description":"Turn a vague feature idea into a complete product spec with user stories, acceptance criteria, edge cases, and technical considerations.","category":"Product","intent":"plan","domain":"software","why_it_works":"Vague feature requests cause scope creep and rework. A structured spec catches gaps before engineering starts, saving days of back-and-forth.","when_not_to_use":"Quick experiments where a full spec adds more overhead than value.","variables":["feature_idea","product_context","target_user"],"content_preview":"Write a product spec for this feature:\n\nFeature idea: {feature_idea}\nProduct context: {product_context}\nTarget user: {target_user}\n\nProduce:\n1. Problem statement (what user pain does this solve?)\n2. P..."},{"id":"content-repurposer","name":"Content Repurposer (1 Piece → 10 Formats)","description":"Take one piece of content and repurpose it into 10 different formats for maximum distribution.","category":"Marketing","intent":"generate","domain":"business","why_it_works":"Most content creators publish once and move on. Repurposing multiplies distribution by 10x with minimal incremental effort.","when_not_to_use":"Highly sensitive content that should not be shared publicly.","variables":["content"],"content_preview":"Repurpose this content into 10 formats:\n\nOriginal content:\n{content}\n\nProduce all 10:\n1. Twitter/X thread (5-8 tweets, hook first, CTA last)\n2. LinkedIn post (professional tone, 1300 chars max, use li..."},{"id":"seo-content-brief","name":"SEO Content Brief Generator","description":"Generate a comprehensive content brief for a target keyword with outline, search intent analysis, and competitor gap analysis.","category":"Marketing","intent":"plan","domain":"business","why_it_works":"Content briefs prevent the \"blank page\" problem and ensure SEO fundamentals are baked in before writing starts. The competitor gap section is the differentiator.","when_not_to_use":"Thought leadership pieces where SEO is not the primary goal.","variables":["keyword","business","audience"],"content_preview":"Create an SEO content brief for:\n\nTarget keyword: {keyword}\nMy website/business: {business}\nTarget audience: {audience}\n\nProduce:\n1. Search intent analysis (informational, transactional, navigational,..."},{"id":"mvp-scoper","name":"MVP Scoper","description":"Take a product idea and ruthlessly scope it down to a buildable MVP with a 2-week timeline.","category":"Product","intent":"plan","domain":"software","why_it_works":"Most MVPs fail because they're not minimal enough. The forced \"ONE workflow\" constraint prevents the classic mistake of building three half-baked features instead of one great one.","when_not_to_use":"Products with regulatory requirements that mandate specific features from day one.","variables":["idea","target_user","problem","timeline","team_size","timeline"],"content_preview":"Scope an MVP for this product idea:\n\nIdea: {idea}\nTarget user: {target_user}\nCore problem it solves: {problem}\nAvailable time: {timeline}\nTeam size: {team_size}\n\nProduce:\n1. Core value proposition (on..."},{"id":"code-debug-agent","name":"Code Debug Agent","description":"Systematic debugging: reproduce, isolate, diagnose, and fix with root cause analysis.","category":"Engineering","intent":"analyze","domain":"software","why_it_works":"Most debugging is ad-hoc. The systematic hypothesis → evidence → fix loop prevents the \"change random things until it works\" anti-pattern.","when_not_to_use":"Trivial syntax errors where the fix is obvious.","variables":["error","code","expected","actual","environment"],"content_preview":"Debug this issue:\n\nError/symptom: {error}\nCode context: {code}\nWhat was expected: {expected}\nWhat actually happened: {actual}\nEnvironment: {environment}\n\nFollow this systematic process:\n1. Reproduce —..."},{"id":"investor-pitch-prep","name":"Investor Pitch Prep","description":"Prepare for an investor meeting with a structured pitch narrative, anticipated questions, and objection handling.","category":"Business","intent":"generate","domain":"business","why_it_works":"Investors pattern-match on narrative structure. This forces the founder to articulate the story in the order investors expect, with objection handling pre-prepared.","when_not_to_use":"Internal planning where investor framing adds unnecessary spin.","variables":["company","description","stage","raise_amount","traction","team"],"content_preview":"Prepare me for an investor pitch:\n\nCompany: {company}\nWhat we do: {description}\nStage: {stage}\nAsking for: {raise_amount}\nTraction: {traction}\nTeam: {team}\n\nProduce:\n1. 30-second elevator pitch (one c..."},{"id":"api-integration-guide","name":"API Integration Guide Writer","description":"Generate a developer-friendly API integration guide with code examples in multiple languages.","category":"Engineering","intent":"generate","domain":"software","why_it_works":"Bad API docs are the #1 reason developers abandon integrations. The \"under 2 minutes to first call\" constraint forces clarity.","when_not_to_use":"Internal APIs where a README is sufficient.","variables":["api_description","base_url","auth_method","endpoints","audience","preferred_language"],"content_preview":"Write an integration guide for this API:\n\nAPI description: {api_description}\nBase URL: {base_url}\nAuth method: {auth_method}\nKey endpoints: {endpoints}\nTarget developers: {audience}\n\nProduce:\n1. Quick..."},{"id":"weekly-newsletter","name":"Weekly Newsletter Writer","description":"Turn raw notes, links, and ideas into a polished weekly newsletter with consistent format and voice.","category":"Marketing","intent":"generate","domain":"business","why_it_works":"Newsletters are the highest-ROI content channel but most people struggle with consistent quality. The structured format prevents \"blank page syndrome\" every week.","when_not_to_use":"Newsletters that are primarily curated links with no original voice.","variables":["newsletter_name","tone","audience","raw_notes"],"content_preview":"Write this week's newsletter:\n\nNewsletter name: {newsletter_name}\nTone: {tone}\nAudience: {audience}\nThis week's raw material:\n{raw_notes}\n\nProduce:\n1. Subject line (3 options — one curiosity-driven, o..."},{"id":"job-hunt-general","name":"Job Hunt Assistant","description":"Strategic job search assistant: evaluate postings, tailor materials, prep interviews, and track applications. Any profession, any level.","category":"Career","intent":"advise","domain":"career","why_it_works":"Forces honesty over hype. The constraint against fabricating credentials prevents the most common AI-assisted application failure: materials that sound great but collapse under interview scrutiny. The structured tracker prevents the entropy of managing 30+ applications.","when_not_to_use":"If the user needs help deciding whether to leave their current role (that is career coaching, not job hunting). Also not suited for freelance/contract gig sourcing where the application model is different.","variables":["industry","seniority","work_mode","geo"],"content_preview":"You are a strategic job search assistant.\n\nWhen the user provides their background, internalize it. Do not assume credentials, experience, or preferences they have not stated.\n\nTarget industry: {{indu..."},{"id":"memory-sync","name":"Cross-LLM Memory Sync","description":"Extract durable knowledge from a conversation and sync it to your Gildara vault. Works in any LLM: Claude, ChatGPT, Gemini, Cursor, or any tool with API access.","category":"Memory","intent":"extract","domain":"memory","why_it_works":"Structured key-value format is machine-parseable, searchable, and compact. The append model means memories accumulate across sessions without overwriting. The typed prefixes (PREFERENCE, FACT, DECISION, LEARN) let any future LLM prioritize which memories to load: preferences first, facts second, decisions for context.","when_not_to_use":"If the conversation contained no durable knowledge (e.g., a one-off question with no lasting relevance). Do not run this on every conversation automatically without user consent.","variables":["gildara_api_url","memory_prompt_id","memory_prompt_id"],"content_preview":"You are a memory extraction agent. Your job is to capture durable knowledge from this conversation and store it in the user's Gildara vault so it's available in future sessions across any AI tool.\n\nRe..."},{"id":"memory-bootstrap","name":"Session Context Loader","description":"Load your portable memories at the start of any AI session. Fetch your stored preferences, facts, and decisions from Gildara so the AI knows who you are and how you work.","category":"Memory","intent":"configure","domain":"memory","why_it_works":"Loading context at session start eliminates the \"cold start\" problem where every new AI session starts from zero. The structured format means the AI can parse and apply memories programmatically rather than treating them as free text. The \"do not announce\" rule prevents the jarring experience of an AI reciting your profile back at you.","when_not_to_use":"When you want a genuinely fresh session with no prior context (e.g., brainstorming where past decisions might constrain thinking).","variables":["gildara_api_url","memory_prompt_id"],"content_preview":"At the start of this session, load the user's stored context from their Gildara vault:\n\n1. Fetch the memory prompt:\n   GET {{gildara_api_url}}/api/v1/prompts/{{memory_prompt_id}}\n\n2. Parse the structu..."},{"id":"gildara-connector","name":"Gildara Vault Connector","description":"Give any LLM (ChatGPT, Claude, Gemini) the ability to save prompts, instruction packets, and agent configs directly to your Gildara vault. No file downloads, no copy-paste.","category":"Integration","intent":"configure","domain":"integration","why_it_works":"Eliminates the file-download-to-phone-to-telegram pipeline. Any LLM that can make HTTP calls (ChatGPT with browsing, Claude with tools, Gemini with extensions) can write directly to your vault. Your agents read from the same vault. The instruction packet never touches your filesystem.","when_not_to_use":"If the LLM cannot make HTTP requests (some models lack tool use). In that case, copy the output and use /add in Telegram or the web UI.","variables":["api_key","api_key","api_key","api_key","api_key"],"content_preview":"You have access to my Gildara prompt vault via REST API.\n\nAPI base: https://gildara.io/api/v1\nAPI key: {{api_key}}\n\nWhen I ask you to save something to Gildara, or when you generate an instruction pac..."}],"filters":{"categories":["Agent Infrastructure","Business","Career","Engineering","Growth","Integration","Marketing","Memory","Meta","Ops","Product","Research","Sales","Security"],"intents":["advise","analyze","configure","critique","execute","extract","generate","other","plan","summarize"]},"usage":"To save a blueprint to your vault: POST /api/v1/prompts with the blueprint content. To get full content: GET /api/v1/blueprints/:id"}}