| name | session-log |
| description | Create a timestamped research-session progress log that records completed work, evidence, decisions, and next steps. Use when closing or checkpointing substantive project work that needs a durable chronology. Not for saving one reusable finding; use $store-insight. |
| allowed-tools | Read, Write, Edit, Bash(mkdir*), Bash(ls*) |
| argument-hint | [project-name-or-path] |
| skill-dependencies | ["update-focus"] |
Session Log Skill
Automatically create timestamped progress logs for research sessions.
Purpose
Based on Scott Cunningham's session workflow: “Progress logs are my autosave of the workflow.” When sessions end or crash, the next AI client can read logs and pick up exactly where you left off.
When to Use
At the end of any significant work session, or when asked to "log this session" or "update progress".
Workflow
Step 1: Identify Projects Touched
Before writing anything, inventory which projects were affected during this session. A "project" is any directory with its own CLAUDE.md or log/ directory. Common splits:
| Scope | Where the log goes |
|---|
| Work inside a specific project | That project's log/ |
| Global infrastructure (skills, hooks, rules, settings) | Task Management's log/ |
| Course/module-level changes (reorganisation, new CLAUDE.md) | That module's log/ |
Signs of a multi-project session:
- Files changed in AI client infrastructure (skills, hooks, settings) → global/infrastructure log
- Files changed in the CWD project → project-specific log
- Files changed in a parent or sibling directory → check if that's a separate project
Step 2: Create One Log Per Project
For each project identified in Step 1:
- Read existing context — check that project's
.context/current-focus.md and recent log/ entries
- Create
log/ directory if it doesn't exist
- Write the log to
log/YYYY-MM-DD-HHMM.md within that project
- Scope the content — each log only covers what happened in that project, not the whole session
If there's only one project, this reduces to a single log (the common case).
Step 3: Cross-Reference
When multiple logs are created, add a brief cross-reference at the top of each:
> Also logged: [other project name] — `[relative or absolute path to other log]`
This lets a future session in one project discover that related work happened elsewhere.
Step 4: Offer Follow-Up
- Offer to run
update-focus for a structured update (session rotation, open loops), rather than making ad-hoc edits to current-focus.md
- If multiple projects were touched, offer to update focus for each
Log Template
# Session Log: [Date] [Time]
## Project: [Project Name]
## What We Did
- [Bullet points of accomplishments]
## Key Decisions
- [Any choices made and why]
## Problems/Blockers
- [Issues encountered]
## Next Steps
- [ ] [Actionable next items]
## Files Changed
- [List of modified files — only those in THIS project]
Examples
Single-project session
"Please log this session — we worked on the research paper, fixed the simulation code, and decided to target Journal B instead of Journal A."
→ One log in the MCDM project's log/
Multi-project session
"Please log this session — we did Workshop 17, reorganised the module folder, and created two new global skills."
→ Three logs:
- Module project
log/ — workshop completion, reorganisation, CLAUDE.md creation
- Task Management
log/ — new skills (init-project-course, audit-project-course), new hook (ensure-latexmkrc.sh), settings.json changes
