agentleFS
benororbenoror/obsidianos_work/.agents/skills/followup-todos/SKILL.md

followup-todos

Skill166 starsChanged 5 months agoMITView on GitHub

Extract action items as plain markdown bullets (with confirmation). Args: <path>. No args = run [/note-status pending --step=todos](../note-status/SKILL.md#pending-mode).

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 4 are distinctive, and 8 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 8

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read the entire file. Also read Tracker.md to cross-reference Jira tickets and current task status — this helps:

    line 34
  2. Distinctive:

    Avoid duplicates: If a follow-up already exists in the Tracker for the same Jira ticket, flag it in the proposal table (Step 3) rather than …

    line 36
  3. Very specific:

    Add context: If the meeting references a Jira ticket ID (e.g.

    line 37
  4. Very specific:

    Default ✅ for high-relevance items owned by the user.

    line 81
  5. Very specific:

    Default ⬜ for others' items or skippable ones.

    line 82
  6. Very specific:

    Do not: use - [ ]. Placement depends on the note's template:

    line 95
  7. Very specific:

    Prefer a dedicated section heading if missing, e.g.

    line 117
  8. Very specific:

    After inserting, add a TodosExtracted: YYYY-MM-DDTHH:MM:SS±HH:MM frontmatter property to prevent re-processing.

    line 126
  9. Distinctive:

    Always read the file before editing — frontmatter may have changed since your last read.

    line 134
  10. Distinctive:

    Do NOT duplicate follow-ups that already exist in the note body (compare by description similarity).

    line 135
  11. Very specific:

    When in doubt about relevance, include it as ⬜ in Step 3 and let the user decide.

    line 136
  12. Distinctive:

    If the meeting is old (>2 weeks), flag items that may already be completed and suggest skipping them.

    line 137

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Follow-up Todos

Usage

  • /followup-todos <path> — Extract and propose follow-ups from a specific meeting note.
  • /followup-todos (no args) — Run /note-status pending --step=todos to show meetings missing follow-up extraction. The user picks one (or more) to process.

Output format (mandatory)

  • Never insert Obsidian Tasks / checkbox lines: no - [ ], - [x], - [-], or any - [?] pattern.
  • Never append Tasks-plugin inline metadata on inserted lines: no 📅, 🛫, ⏳, or other parsed task tokens. (Optional priority emoji at the very end as plain text — e.g. 🔼 — is fine if it reads naturally in prose.)
  • Always insert plain list items: - Description… Wikilinks ([[@Name]], [[+Team]]) and bold are allowed.

Prerequisites

The meeting file should ideally have cached AI transcripts (from /cache-notes). If not cached, suggest running /cache-notes <path> first. However, the skill also works on notes with only manual content.

Optional context

For classification (owner, urgency, dates in the proposal step), you may still use people-resolver and obsidian-tasks as reference — but do not copy Tasks checkbox or date-token syntax into the note.

Workflow

Step 1: Read the meeting note and gather context

See vault-context for vault discovery conventions.

Read the entire file. Also read Tracker.md to cross-reference Jira tickets and current task status — this helps:

  • Set priority: A ticket already marked 🔺 in the Tracker should keep that priority (as trailing emoji in the bullet text if you include one).
  • Avoid duplicates: If a follow-up already exists in the Tracker for the same Jira ticket, flag it in the proposal table (Step 3) rather than creating a duplicate line in the note.
  • Add context: If the meeting references a Jira ticket ID (e.g. PROJ-1234), pull its current status and assignee from the Tracker.

Extract action items from all content sources, skipping lines that are already captured as follow-ups (compare meaning, not only exact text):

  1. Manual notes (between frontmatter --- and ## 🤖 AI Notes): Free-text action items (e.g. "- Talk to Alex about X"). Skip lines that are checkbox tasks (- [*] — any single character between brackets) when deduplicating, but note the vault may still contain legacy - [ ] lines from older runs.
  2. [!gemini_todos] callout: Gemini's "Suggested Next Steps" — explicit action items.
  3. [!gemini_notes] callout: Summary & Details — scan for implicit commitments ("will do X", "agreed to Y", "plans to Z").
  4. Other provider callouts ([!otter_todos], etc.): Same treatment as Gemini.

Also extract:

  • Existing plain bullets and checkbox lines in the follow-up section to avoid duplicates
  • Participants: from frontmatter (to resolve assignees)
  • Meeting date (from filename or created: frontmatter)

Step 2: Classify each action item

For each candidate action item (from manual notes, transcript todos, or implicit commitments in details), determine:

  1. Owner: Who is responsible? See people-resolver for name matching and assignee rules.
  2. Relevance: Score as high, medium, or skip:
    • High: Action is for the user, has a clear deliverable, or is time-sensitive.
    • Medium: Action is for someone else but the user should track it, or it's vague but potentially important.
    • Skip: Purely informational, already completed (based on date vs today), or not actionable.
  3. Priority (for the proposal table only): Based on urgency and impact (see priorities in obsidian-tasks). When inserting bullets, you may end the line with a single priority emoji (🔺 ⏫ 🔼 🔽) as plain text — never as Tasks metadata.
  4. Dates: Note due/start in the proposal table when helpful; do not put 📅 / 🛫 / ⏳ on inserted lines. If a date matters, write it in words in the bullet (e.g. "by 2026-03-26").

Step 3: Present proposals to the user — MANDATORY CONFIRMATION

⚠️ STOP HERE AND WAIT FOR USER CONFIRMATION. Never skip this step, even during /meeting wrap sequences. Do NOT write follow-ups to the file until the user explicitly approves.

Display a numbered table:

| # | Add? | Follow-up | Owner | Priority |
|---|------|-----------|-------|----------|
| 1 | ✅   | Talk with Alex about temp environments | Me | 🔼 |
| 2 | ✅   | Offload data lake work to Jane | Me | ⏫ |
| 3 | ⬜   | Confirm travel plans for Friday | [[@Sam Rivera]] | — |
| 4 | ⬜   | Verify production deploy completed | Me | 🔼 |
  • Default ✅ for high-relevance items owned by the user.
  • Default ⬜ for others' items or skippable ones.
  • Let the user toggle by saying numbers (e.g. "1,3,4" or "all" or "none except 2").

Step 4: Insert confirmed follow-ups

Only proceed after the user has explicitly confirmed which items to include (Step 3).

Insert each confirmed item as one line:

- Clear, self-contained description (optional **bold**, `[[@Name]]`, optional trailing 🔼)

Do not use - [ ]. Placement depends on the note's template:

Daily Standup notes (Meetings/*/Scrum/YYYY-MM-DD.md)

These follow the Daily Standup template:

Yesterday
- ...
Today
- ...          ← INSERT FOLLOW-UPS HERE (append after existing Today items)
Blockers
- ...
---
Pending/Carry-over Backlog
- ...

Insert at the end of the "Today" section, just before the Blockers line. Use StrReplace targeting Blockers as the anchor.

Other meeting notes

Prefer a dedicated section heading if missing, e.g. ## Follow-ups, placed after the title / any manual notes and before ## 🤖 AI Notes. If a ## Follow-ups (or equivalent) section already exists, append there. If neither heading exists, append after the frontmatter closing --- and add ## Follow-ups first.

General rules

  • Group by owner if multiple people are involved (optional sub-bullets under a short **[[@Name]]** line).
  • Preserve existing content — always append, never overwrite.

Step 5: Mark as processed

After inserting, add a TodosExtracted: YYYY-MM-DDTHH:MM:SS±HH:MM frontmatter property to prevent re-processing. This is required — /meeting wrap pending relies on it to detect unprocessed meetings.

Step 6: Offer to commit

See /commit. Skip when called as part of a sequence (e.g. /meeting wrap).

Important Notes

  • Always read the file before editing — frontmatter may have changed since your last read.
  • Do NOT duplicate follow-ups that already exist in the note body (compare by description similarity).
  • When in doubt about relevance, include it as ⬜ in Step 3 and let the user decide.
  • If the meeting is old (>2 weeks), flag items that may already be completed and suggest skipping them.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.