devlog-harvester
Skill2 starsChanged 10 months agoMITView on GitHub
Automatically harvest session data, git activity, and project context. Smart triggers: git commits, project switches, time-based (3x daily), idle detection. Feeds into knowledge-forge for Obsidian storage.
In this repository17 other files it gives its agents.
- react-best-practices/AGENTS.mdAGENTS.md
- agent-browserSkill
- claudeceptionSkill
- humanizerSkill
- knowledge-forgeSkill
- loop-maestroSkill
- notebooklmSkill
- obsidian-bridgeSkill
- portfolio-forgeSkill
- project-analyzerSkill
- pr-review-resolverSkill
- vercel-react-best-practicesSkill
- vaultage-captureSkill
- vaultage-dailySkill
- vaultage-inboxSkill
- vaultage-searchSkill
- web-design-guidelinesSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 4 are distinctive, and 0 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.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 28
Every commit triggers a mini-harvest capturing:
- Distinctive: line 39
When changing project directories:
- Distinctive: line 44
After 30 minutes of no tool calls:
- Distinctive: line 111
After 30 minutes of inactivity
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.
Files
DevLog Harvester
Captures development session data through smart, non-intrusive triggers.
Trigger Mechanisms
1. Git Commit Hook (Primary)
Every commit triggers a mini-harvest capturing:
- Files changed in commit
- Commit message (intent)
- Session context at commit time
2. Time-Based Rollups (3x Daily)
Scheduled at: 9:00, 14:00, 21:00
- Aggregates accumulated data
- Writes to Obsidian via knowledge-forge
3. Project Switch Detection
When changing project directories:
- Flush current project's accumulated data
- Initialize harvester for new project
4. Idle Flush (30 min)
After 30 minutes of no tool calls:
- Flush accumulated session data
- Prevents data loss
Data Captured
Per-Session Data
session:
project: ATLAS-main
started: 2026-01-19T09:15:00
ended: 2026-01-19T12:30:00
duration_minutes: 195
files_touched:
- path: src/api/client.py
operations: [read, edit, read]
lines_changed: +45/-12
- path: tests/test_api.py
operations: [create, edit]
lines_changed: +120/-0
git_activity:
commits: 3
messages:
- "feat(api): add chat endpoint"
- "test: add API client tests"
- "fix: handle timeout errors"
total_insertions: 165
total_deletions: 12
tools_used:
Edit: 12
Read: 45
Bash: 23
Grep: 8
errors_encountered:
- type: "ImportError"
file: "src/api/client.py"
resolution: "Added missing httpx dependency"
decisions_made:
- "Used httpx instead of requests for async support"
- "Added retry logic with exponential backoff"Usage
Manual Harvest
# Harvest current session
/harvest
# Harvest specific project
/harvest --project ATLAS-main
# Harvest and immediately write to Obsidian
/harvest --writeAutomatic (via hooks)
The harvester runs automatically when:
- You make a git commit
- You switch projects (cd to different repo)
- 3x daily at scheduled times
- After 30 minutes of inactivity
Integration Points
Input Sources
- Git history (
git log,git diff) - Session transcripts (
~/.claude/transcripts/) - Tool usage (from session context)
- AGI-memory recalls
Output Destinations
- →
knowledge-forge(for synthesis) - →
obsidian-bridge(for storage) - → Local cache (
~/.cache/devlog/)
Data Flow
┌─────────────┐ ┌─────────────┐ ┌─────────────┐
│ Triggers │────▶│ Harvester │────▶│ Cache │
│ (various) │ │ Engine │ │ (staging) │
└─────────────┘ └─────────────┘ └──────┬──────┘
│
On rollup/flush │
▼
┌─────────────┐
│ knowledge │
│ forge │
└─────────────┘Configuration
Set in ~/.config/devlog/config.yaml:
harvester:
# Rollup times (24h format)
rollup_times: ["09:00", "14:00", "21:00"]
# Idle threshold before flush (minutes)
idle_threshold: 30
# Cache directory
cache_dir: ~/.cache/devlog
# Projects to track (empty = all)
tracked_projects: []
# Projects to ignore
ignored_projects:
- node_modules
- .gitFiles
scripts/harvest.py- Main harvester enginescripts/git-post-commit-hook.sh- Git hook installerscripts/scheduler.py- Time-based trigger daemon
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
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.

