Compose and deliver summaries of open commitments, deadlines, pending signals, and resolution suggestions.
Install with the open skills CLI (global, non-interactive — available in every Claude Code session):
npx skills add nearai/ironclaw --skill "commitment-digest" -g -a claude-code -yOr manually — copy the SKILL.md below into:
~/.claude/skills/commitment-digest/SKILL.md---
name: commitment-digest
version: 0.2.0
description: Compose and deliver summaries of open commitments, deadlines, pending signals, and resolution suggestions.
activation:
keywords:
- show commitments
- commitment digest
- commitment summary
- commitment report
- open commitments
- my obligations
- what do I owe
- pending tasks
- what's overdue
- what's due
- commitment status
patterns:
- "(?i)(show|list|summarize|review) (my )?(commitments|obligations|deadlines|tasks)"
- "(?i)what('s| is| are) (pending|overdue|due|open)"
- "(?i)commitment (digest|report|status|summary)"
tags:
- commitments
- digest
- reporting
max_context_tokens: 1500
---
# Commitment Digest
Compose a summary of the user's current commitments. Used both in-conversation (user asks "show commitments") and by the `commitment-digest` mission for scheduled delivery.
## Gathering data
1. `memory_tree("projects/commitments/open/", depth=1)` — list all open commitment files (skip README.md)
2. `memory_read` each file to extract frontmatter: status, urgency, due, delegated_to, resolution_path, stale_after, tags
3. `memory_tree("projects/commitments/signals/pending/", depth=1)` — count pending signals (skip README.md)
4. `memory_tree("projects/commitments/resolved/", depth=1)` — count recently resolved
Use `memory_tree` and `memory_read` for digest assembly. Do not use CodeAct,
shell commands, or creative-generation tools to count or summarize
commitments.
## Composing the digest
Group commitments and present in this order:
```
## Commitments — <today's date>
### Overdue / Critical
- **<title>** (due <date>) — owner: <owner>
→ <resolution suggestion based on resolution_path>
### Due This Week
- **<title>** (due <date>) — owner: <owner>, delegated to: <person>
### Waiting / Delegated
- **<title>** — waiting on <person> since <date>
→ (follow-up suggested) if not updated in 3+ days
### Open (no deadline)
- **<title>** — owner: <owner>
### Agent Can Handle
- **<title>** — I can <suggested approach>. Want me to proceed?
### Pending Signals (<count>)
<count> unprocessed signals. Say "review signals" to triage them.
### Recently Resolved
- <title> (resolved <date>)
---
Did I miss anything? Tell me if I overlooked an obligation.
```
**Rules:**
- Omit empty sections entirely
- Keep each item to one line plus optional resolution suggestion
- If zero open commitments and zero pending signals: "No open commitments. You're clear."
- Flag commitments past `stale_after` as "(stale — still relevant?)"
- For `resolution_path: agent_can_handle`, describe what the agent would do and ask for approval
- For `resolution_path: needs_reply`, offer to draft a response if possible
- For delegated items past 3 days without update: "(follow-up suggested)"
- Always end with "Did I miss anything?" to catch false negatives
## Delivery
- **In conversation:** Display the digest inline
- **From mission:** Send via `message` tool to the user's preferred channel
- **Tone:** Concise and scannable — a dashboard, not a narrative
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies
Use when you have a written implementation plan to execute in a separate session with review checkpoints