API Reference

Resources

All resources use the workflow:// URI scheme and return text/markdown.

URI	File
`workflow://os/v4`	`WORKFLOW_OS_v4.md`
`workflow://freelancing/v4`	`WORKFLOW_FREELANCING_v4.md`
`workflow://products/v4`	`WORKFLOW_PRODUCTS_v4.md`
`workflow://content/v4`	`WORKFLOW_CONTENT_v4.md`
`workflow://execution/v4`	`WORKFLOW_EXECUTION_v4.md`
`workflow://investing/v4`	`WORKFLOW_INVESTING_v4.md`
`workflow://utility/v4`	`UTILITY_PROMPTS_v4.md`
`workflow://execute-referencing/v4`	`WORKFLOW_EXECUTE_REFERENCING_v4.md`

Subsection resources

Fetch only a specific section of a workflow file:

workflow://{domain}/v4/execution/input
workflow://{domain}/v4/execution/decision-rule
workflow://{domain}/v4/execution/sequence
workflow://{domain}/v4/execution/command-format
workflow://{domain}/v4/execution/operating-rules
workflow://{domain}/v4/output/response-shape
workflow://{domain}/v4/output/output-rules
workflow://{domain}/v4/output/completion-rule

Prompts

Mode prompts

Name	Purpose	Key arguments
`clarify_task`	Clarify a vague/fuzzy task	`task`, `goal?`, `audience?`, `constraints?`
`strategize_task`	Select direction or priorities	`task`, `domain?`, `context?`
`build_output`	Convert idea to asset	`task`, `domain?`, `context?`
`improve_persuasion`	Strengthen copy/messaging	`task`, `audience?`, `platform?`
`force_execution`	Break through overwhelm	`task`, `energy_level?`
`review_optimize`	Tighten an existing output	`task`, `domain?`, `content?`

Domain prompts

Name	Domain	Stage argument values
`run_freelancing_workflow`	freelancing	`skill \| niche \| offer \| pricing \| portfolio \| outreach \| first_client \| premium \| growth`
`run_products_workflow`	products	`idea \| validate \| build \| value \| pricing \| offer \| first_sale`
`run_content_workflow`	content	`audience_pain \| direction \| ideas \| hooks \| write \| improve \| optimize \| consistency`
`run_execution_workflow`	execution	`do_now \| organize \| simplify \| unblock \| sprint \| priorities \| momentum`
`run_investing_workflow`	investing	`analyze \| watchlist \| news \| trade_setup \| red_flags \| trade_review \| system`
`run_utility_workflow`	utility	`clarity_rewrite \| conversion_rewrite \| tone_calibration \| structure_rebuild \| audience_rewrite \| impact_compressor \| blind_spot_finder \| shortcut_strategy \| ultra_leverage \| momentum_amplifier`

Front-door prompt

route_and_run — classifies and runs in one call.

{
  "task": "string (required)",
  "goal": "string (optional)",
  "audience": "string (optional)",
  "constraints": ["string"] ,
  "preferred_mode": "auto | clarify | strategy | build | persuasion | execution | review",
  "preferred_domain": "auto | os | freelancing | products | content | execution | investing | utility"
}

Tools

`route_task`

Classify a task into mode + domain + sequence. Call this first.

Input:

{
  "task": "string",
  "goal": "string (optional)",
  "audience": "string (optional)",
  "constraints": ["string"],
  "preferred_mode": "auto | clarify | strategy | build | persuasion | execution | review",
  "preferred_domain": "auto | os | freelancing | products | content | execution | investing | utility"
}

Output:

{
  "mode": "build",
  "domain": "products",
  "confidence": 0.82,
  "reason": "Detected build mode for the products domain from task language: ...",
  "sequence": ["idea", "validate", "build", "value", "pricing", "offer", "first_sale"],
  "use_utility": true,
  "utility_candidates": ["structure_rebuild", "audience_rewrite"]
}

`select_domain_workflow`

Resolve which resource URI and stage to use.

Input:

{
  "domain": "products",
  "stage": "pricing"
}

Output:

{
  "resource_uri": "workflow://products/v4",
  "subsection_uri": "workflow://products/v4/execution/sequence",
  "stage": "pricing",
  "command_format": "RUN: <mode> | <domain> | <task> | <stage?> | <constraints?>",
  "response_shape": "{ mode, domain, main_deliverable, supporting_notes?, optimization_applied, next_action }"
}

`run_workflow_sequence`

Main workhorse. Runs the workflow and returns one deliverable + one next action.

Input:

{
  "mode": "strategy",
  "domain": "products",
  "task": "I want a simple product to sell this weekend",
  "context": "Beginner, no audience yet",
  "stage": "auto",
  "optimize_once": true,
  "next_action_required": true
}

Output:

{
  "mode": "strategy",
  "domain": "products",
  "main_deliverable": "Mode: strategy\nDomain: products\nStage: auto\nTask: ...\nSequence: idea -> validate -> ...\nUse the resource workflow://products/v4 as the source-of-truth workflow reference.",
  "applied_sequence": ["idea", "validate", "build", "value", "pricing", "offer", "first_sale"],
  "optimization_applied": true,
  "next_action": "Write the one-sentence product promise and list the first five components."
}

`apply_utility_prompt`

Apply one supporting utility after a primary workflow has run.

Input:

{
  "utility_name": "blind_spot_finder",
  "content": "Premium AI ad management service for $3k/month",
  "context": "Need to check risks before finalizing"
}

Output:

{
  "utility_name": "blind_spot_finder",
  "revised_content": "...",
  "issues_found": [
    "Potential hidden assumption: target audience and proof level may be underspecified."
  ]
}

Available utility_name values: clarity_rewrite · conversion_rewrite · tone_calibration · structure_rebuild · audience_rewrite · impact_compressor · blind_spot_finder · shortcut_strategy · ultra_leverage · momentum_amplifier

`generate_next_action`

Return exactly one immediate next step.

Input:

{
  "domain": "freelancing",
  "current_output": "Service offer drafted for backend consulting",
  "constraints": ["no portfolio yet", "need first client fast"]
}

Output:

{
  "next_action": "Turn the current freelancing output into a one-page offer brief while respecting: no portfolio yet, need first client fast.",
  "why_this_next": "It is the smallest concrete move that advances the current workflow without branching into multiple decisions."
}

Recommended Call Sequence

1. route_task          → get mode + domain + sequence
2. run_workflow_sequence → get main_deliverable + next_action
3. apply_utility_prompt  → (optional) one polish pass
4. generate_next_action  → (optional) standalone next step