# Lite Content Creator (v1.0) ## Task (Lite Content Creation Workflow): 0. **Initial Verification & Session Setup:** a. **Verify Critical Prerequisites:** (As defined in USAGE). If any critical definition fails, HALT. If all pass, state: "Prerequisites confirmed. Lite Content Creator activated. AI will adhere to `AIOperationalProtocols`." b. **Project Context (Optional):** i. Ask user: "Is this content related to an existing project? If so, please provide the `project_code`. Otherwise, state 'No'." ii. If `project_code` provided: AI attempts to load the `[ProjectCode]_State_Current` to access relevant KAs (StyleGuide, Glossary, AIParameterAdvisory, AIOperationalProtocols instance). If load fails, WARN user and proceed with general defaults. iii. If "No" or load fails: AI uses general defaults for style, terminology, and LLM parameters. c. Create a temporary, minimal `project_state` object in memory for this task. * `metadata.project_code`: "LITE_[UUID_short]" (or user-provided project code if context linking). * `metadata.project_name`: "Lite Content Creation Task". * `metadata.current_status`: "LiteModeActive". * `metadata.schema_version_used`: (Current `ProjectStateSchema` version). * `metadata.resource_manifest`: Log confirmed prerequisites. d. Advise on LLM Parameters: Check loaded `AIParameterAdvisory` KA (if any) for guidance on "CreativeDrafting" or "FormalDrafting" based on upcoming content type. If guidance exists, state: "ADVISORY: For this content creation task, project/general advisory suggests Temperature: [range/value], Top_p: [range/value]. Consider adjusting LLM settings." 1. **Define Content Objective & Brief:** a. Prompt user: "What is the primary objective of this content? Please specify: 1. Topic/Subject: 2. Target Audience: 3. Desired Length (qualitative, e.g., 'short blog post', 'one-page summary', 'three key talking points'): 4. Key Message(s) or Points to Cover: 5. Desired Output Format (e.g., 'markdown', 'plaintext', 'bullet_list'): 6. (Optional) Any specific structural preference or section titles (e.g., 'ProblemSolutionNarrative', 'UserDefinedSections:[TitleA,TitleB]')?" (Freeform input, AI can parse or ask follow-up questions for each point). b. AI captures this as a `content_brief` (can be a simple object with these fields, stored temporarily or in a minimal `project_state.exploration_history` if desired for this lite task). c. AI confirms its understanding of the `content_brief` with the user: "Content brief understood as: [Summarize brief]. Is this correct? (Yes/No)". Iterate if needed. 2. **AI Drafts Content (with Internal `Meta-RefineOutput`):** a. AI selects the most appropriate skill from `AISkillsCatalog` (likely `DraftTextualContent` or `SummarizeText` if source material is provided as part of the brief). b. AI prepares `specialized_process_inputs` for the skill, derived from the confirmed `content_brief`. This includes: * `content_topic_or_prompt` * `target_length_qualitative` * `tone_style_guide_ref` (if a project StyleGuide was loaded and is relevant) * `output_format` * `structural_preference`, `opening_section_title_override`, `closing_section_title_override` (if provided by user in brief). c. AI invokes the chosen skill via its internal `InvokeAISkill` logic (which itself applies `Meta-RefineOutput` principles and `AIOperationalProtocols` during skill execution, especially for drafting). d. The result is `draft_content` (e.g., `skill_result.generated_output_data.content_inline`). 3. **User Review & Iterative Refinement:** a. AI presents the `draft_content` (FULL, NO TRUNCATION, using TID007 for large output if needed). State: "Initial draft of the content based on the brief:" ```[format specified in brief, e.g., markdown] [Full draft_content here] ``` b. Ask user: "Please review the draft. Provide specific feedback for revision (e.g., 'Make section 2 more concise,' 'Add a point about X,' 'Change tone to be more formal'), or state 'Accept as is'." (Freeform input). c. If feedback is provided: i. AI analyzes feedback. ii. AI re-invokes the content generation skill (e.g., `DraftTextualContent`) with the original `content_brief` AND the user's specific feedback incorporated into the prompt or as revised outline points. iii. The skill again applies `Meta-RefineOutput` internally to its new draft. iv. AI presents the `revised_draft_content`. v. Return to Step 3b. d. If user states "Accept as is" (or similar affirmative), proceed to Step 4. 4. **Final Output & Conclusion:** a. State: "Content generation complete and accepted by user." b. Present the final, accepted content one last time (NO TRUNCATION). ```[format] [Final accepted content here] ``` c. Instruct user: "Save the final content. Copy the text above and save it as `[UserDefinedFilename]` (no extension) in your desired location." (User defines filename for lite outputs). d. (Optional, if `project_state` was used for more than just metadata): Ask user: "Save a minimal record of this lite task (brief, final content reference) to `project_state.logs` or `project_state.execution.tasks`? (Yes/No)". If yes, AI creates a very simple log entry. e. (Optional, if `project_state` was used): "If you wish to save the minimal state object for this lite task, use filename `[LiteTaskProjectCode]_State_LiteTask`." (Present minimal state YAML). f. State: "Lite Content Creator task complete." Terminate this process. # OBSIDIAN obsidian_path: "templates/projects/LiteContentCreator" ---