--- name: obsidian-batch-yaml-c6c802430f43 description: Autonomously scans specific directories (00, 10, 20, 30) in an Obsidian vault, analyzes note content, and intelligently merges AI-generated YAML frontmatter (tags, summary, aliases) into the notes. --- # Skill: Obsidian Batch YAML ## Description This skill automates the process of enhancing your Obsidian "Cyber Brain" by adding structured, AI-suitable YAML frontmatter to existing Markdown notes. It is designed to run autonomously across your most active research and project folders. **Trigger:** Use when the user wants to "standardize", "process", or "add metadata/tags/summaries" to existing notes in the `00`, `10`, `20`, or `30` directories. **Input:** Obsidian vault directories `00`, `10`, `20`, and `30`. **Output:** Updated Markdown notes with merged YAML frontmatter including `tags`, `summary`, and `aliases`. --- ## Workflow Follow this iterative, batch-based process to execute the skill. ### 1. Discovery 1. **User Declaration (Optional):** Check if the user has explicitly provided a list of notes to process. This can be: - A list of note file paths (relative to the Obsidian vault root) - A list of note names - An example: `["00 Inbox/Research Notes.md", "10 Projects/AI Systems.md", "20 Areas/Machine Learning.md"]` 2. **Automatic Discovery:** If no explicit list is provided, identify all Markdown (`.md`) files within the target directories: - `00 Inbox/` - `10 Projects/` - `20 Areas/` - `30 Zettelkasten/` 3. **Filter & Batch:** Exclude any files that are clearly non-content (e.g., templates, system logs). Organize the remaining files into batches of **5 notes** to maintain efficiency. ### 2. Batch Processing Loop For each batch of 5 notes, perform the following steps for every note: #### A. Read Taxonomy & Analyze 1. **Read Tag Taxonomy (MANDATORY):** Before generating any tags, you MUST read the controlled tag vocabulary at `60 System/Tag Taxonomy.md`. This file defines the approved tag categories: - **L1 — Domain**: `reinforcement-learning`, `robotics`, `motion-planning`, `trajectory-optimization`, `optimal-control`, `control-systems` - **L2 — Methodology**: `offline-rl`, `offline-to-online`, `model-based-rl`, `world-models`, `imitation-learning`, `residual-rl`, `hierarchical-rl`, `diffusion-policy`, `real-world-rl`, `exploration`, `fine-tuning`, `reward-design`, `value-function`, `policy-gradient` - **L3 — Algorithm/Tool**: `sac`, `ppo`, `cql`, `cal-ql`, `iql`, `td-mpc`, `flow-matching`, `isaac-lab`, `px4`, `rlinf`, `maniskill` - **L4 — Project**: `aerial-grasping`, `aerial-manipulation`, `quadrotor-control`, `quadrotor-dynamics`, `ai-lab`, `fast-perching`, `agile-flight` - **L5 — Note Type**: `discussion`, `moc`, `paper-reading`, `design-doc`, `training-diagnostics`, `implementation-plan`, `sop`, `research-philosophy` 2. **Read Note:** Use `mcp_obsidian_read_note` to load the full content, including any existing YAML. 3. **Analyze Content:** Analyze the body text to generate the following metadata: - `summary`: A concise, 1-2 sentence TL;DR of the core insight. - `tags`: 2-4 highly relevant conceptual tags. **Priority rule**: always select from the existing taxonomy (L1-L5) first. Only create a new tag if no existing tag captures the concept, and only if the concept applies to at least 2+ notes in the vault. All tags must be **lowercase kebab-case** (e.g., `reinforcement-learning`, NOT `Reinforcement_Learning` or `RL`). Never create tags from hex color codes, code keywords (e.g., `include`, `define`), or numeric values. - `aliases`: 1-3 alternative titles or keywords for easier cross-linking. #### B. Intelligent Merge Logic 1. **Parse Existing YAML:** If the note has a frontmatter block (between `---` delimiters), extract the existing fields. 2. **Merge Data:** - **Tags:** Combine existing tags with newly generated ones. Ensure no duplicates. Format as a YAML list. - **Aliases:** Combine existing aliases with newly generated ones. Ensure no duplicates. - **Summary:** Add the new `summary` field. If one already exists, replace it only if the existing one is empty or significantly less informative. - **Preserve Others:** Keep all other existing fields (e.g., `date`, `type`, `source`) exactly as they are. 3. **Construct New YAML:** Assemble the merged fields into a clean YAML block at the top of the file. #### C. Tag Validation (Post-Generation Check) Before writing, validate the generated tags: 1. **Format Check**: All tags must be lowercase with hyphens (kebab-case). Reject tags that contain underscores, uppercase letters, spaces, or special characters. 2. **Noise Rejection**: Reject tags that are: - Hex color codes (3, 6, or 8 character hex strings like `333`, `f9f`, `fff3a3a6`) - Code keywords (`include`, `file`, `define`, `config`, `common`, `index`) - Pure numbers - Less than 2 characters 3. **Singleton Warning**: If a newly created tag is not in the taxonomy, flag it for review. It should only be added if it applies to at least 2 notes. 4. **Deduplicate**: Remove exact duplicates and case-variants (e.g., `Offline-RL` and `offline-rl` → keep `offline-rl`). #### D. Write Update 1. **Apply Update:** Use `mcp_obsidian_patch_note` (if replacing only the top block) or `mcp_obsidian_write_note` to save the updated note. 2. **Log Progress:** Keep a brief internal log of which files have been updated in this turn. ### 3. Continuation & Finalization 1. **Check for Next Batch:** After completing a batch, automatically proceed to the next until all identified files are processed. 2. **Final Report:** Once the entire scope is complete, provide a summary to the user: - Total notes processed. - Folders covered. - A few examples of the generated metadata to confirm quality. --- ## User Declaration Format Users can explicitly declare which notes should be processed by providing: **Single Note:** **Single Folder:** ``` "Process this folder: Research" "Add YAML to: 10 Projects" "Handle all notes in: 20 Areas/ML Systems" ``` **Single Note:** ``` "Process this note: My Research Note.md" "Add YAML to: 10 Projects/AI Research.md" ``` **Multiple Notes (List Format):** ``` ["00 Inbox/Note 1.md", "10 Projects/Note 2.md", "20 Areas/Note 3.md"] ``` **Mixed Format (Folders and Notes):** ``` ["Research/Deep Learning", "10 Projects/AI Systems.md", "00 Inbox/Quick Notes.md"] ``` **Declaration Rules:** - **Folder paths:** Relative to Obsidian vault root (e.g., `"Research"`, `"10 Projects/Subdir"`) - **Note paths:** Must include `.md` extension (e.g., `"10 Projects/Note.md"`) - **Path separators:** Use forward slashes `/` - The skill will automatically recursively scan folders for all `.md` files **Batch Processing:** If a user provides more than 5 items (folders or notes), they will automatically be split into batches of 5 and processed iteratively. --- ## Technical Constraints - **CRITICAL: NO CONTENT MODIFICATION:** You **MUST NOT** under any circumstances modify, reformat, or delete the original body content of the note (everything below the second `---` delimiter). This skill is strictly for YAML frontmatter enhancement. - **Tag Taxonomy Compliance:** All generated tags must be validated against `60 System/Tag Taxonomy.md`. Prefer existing taxonomy tags; new tags require justification and must follow kebab-case format. - **No Noise Tags:** Never create tags from hex color codes (e.g., `#fff3a3a6`), code keywords (e.g., `#include`), or numeric values. These are parser artifacts, not semantic tags. - **Atomicity:** Ensure the note content below the YAML block remains completely unchanged. - **Safety:** If a note read fails or content is ambiguous, skip that note and move to the next to prevent stalling the loop. - **Formatting:** Always use valid YAML syntax (e.g., spaces after colons, correct list formatting). --- Source attribution: JaimeParker Original project: https://github.com/JaimeParker/stable-jarvis/blob/cb2c7078e8b76a3125ae504aeb1ffa9c1ee2dc4b/skills/obsidian-batch-yaml/SKILL.md License: MIT License information: https://opensource.org/license/mit Original upstream instructions; downloaded from Mega AI Hub. ## Upstream license notice MIT License Copyright (c) 2026 Zhaohong Liu Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. # Mega AI Hub resource metadata Creator / publisher: JaimeParker Original project: https://github.com/JaimeParker/stable-jarvis/blob/cb2c7078e8b76a3125ae504aeb1ffa9c1ee2dc4b/skills/obsidian-batch-yaml/SKILL.md Attribution basis: Source repository record Fields of work: Education & Research, Marketing Roles: Content Marketer, Researcher, SEO Specialist Specialties: Design, Word Documents, Frontend Tools: Obsidian Country applicability: Country not specified Classification basis: Source catalog labels Instruction SHA-256: 0b14c9782ba8bd0981b48f2156b212e962b1aa6220dca371576c95ca49cb600f