---

name: project-logger description: Record project activity, decisions, and outputs; query history and status. Provides institutional memory across sessions. triggers:

• log this
• record that
• what happened with
• project history
• what's the status
• מה הסטטוס
• מה קרה עם הפרויקט

---

Project Logger Skill

Maintain a running log of all project activities, decisions, and outputs. Serves as the project's institutional memory and provides queryable history.

Purpose

This skill records all significant project activities, enabling context retention across sessions and providing a queryable history. It supports automatic logging after actions and manual queries for project status.

Modes

Write Mode

• Trigger: After any significant action
• Purpose: Record activities as they happen

Read Mode

• Trigger: Status check, context loading
• Returns: Relevant log entries
• Purpose: Provide historical context

Summarize Mode

• Trigger: Project overview request
• Returns: Condensed project history
• Purpose: Quick status understanding

Log Entry Types

Output Format

Copy# Project Log: {project_name}

**Project Started:** {start_date}
**Last Updated:** {timestamp}

---

## {date}

### {timestamp} - {action_type}
**Actor:** {human|skill_name|external_agent}
**Action:** {what_was_done}
**Output:** {deliverable_or_result}
**Decision:** {if_decision_was_made}
**Next Steps:** {identified_follow_ups}
**Blockers:** {if_any}

---

See resources/log_template.md for full template.

Workflow

Write Mode Workflow

Step 1: Receive Notification

Receive action notification with:

• skill_name: Which skill performed the action
• action_type: One of the 9 LOG_ENTRY_TYPES
• action_details: What was done
• outputs: Any deliverables produced
• next_steps: Follow-up actions identified

Step 2: Format Entry

Structure into log entry format:

• Add timestamp
• Format as markdown section
• Include all relevant fields

Step 3: Append to Log

• Append to {project}_log.md
• Position: Top (newest first)
• Maintain chronological grouping by date

Read Mode Workflow

Step 1: Parse Query

Understand what information is needed:

Step 2: Retrieve Entries

Extract relevant entries from log file.

Step 3: Return Formatted

Return entries in readable format, grouped logically.

Summarize Mode Workflow

Step 1: Analyze Full Log

Process entire log to extract:

• Key milestones reached
• Major decisions made
• Current blockers
• Recent activity

Step 2: Generate Summary

Copy## Project Status Summary: {project_name}

**Health:** {green|yellow|red}
**Progress:** {percent}% toward goals

### Recent Activity (Last 7 days)
- {activity_1}
- {activity_2}

### Key Decisions Made
- {decision_1}

### Current Blockers
- {blocker_1} (since {date})

### Next Milestones
- {milestone}: {target_date}

Health Indicators:

• 🟢 Green: Active progress, no blockers
• 🟡 Yellow: Slow progress or minor blockers
• 🔴 Red: Stalled, major blockers

Example: Auto-Logging

After project-brief creates a brief:

Copy## 2025-12-17

### 18:30:00 - PROJECT_CREATED
**Actor:** project_brief skill
**Action:** Created new project brief
**Output:** ai_strategy_law_faculty_brief.md
**Decision:** N/A
**Next Steps:** Create task breakdown
**Blockers:** None

After delegation-advisor assigns a task:

Copy### 19:15:00 - TASK_DELEGATED
**Actor:** delegation_advisor skill
**Action:** Delegated research task
**Output:** Handoff prompt generated for Gemini
**Decision:** User approved Gemini for research
**Next Steps:** Execute research in Gemini
**Blockers:** None

Example: Status Query

User Request:

"מה הסטטוס של פרויקט AI Strategy?"

Response:

Copy## Project Status Summary: AI Strategy (Law Faculty)

**Health:** 🟡 Yellow
**Progress:** 25% toward goals

### Recent Activity (Last 7 days)
- Created project brief (Dec 17)
- Defined 5 initial tasks (Dec 17)
- Delegated research to Gemini (Dec 17)

### Key Decisions Made
- Chose to focus on governance framework first

### Current Blockers
- Waiting for research results (since Dec 17)

### Next Milestones
- Complete framework research: Dec 20
- Draft proposal for Dean: Dec 25

Cross-Interface Notes

• Claude AI: Log entries created via skill, stored in Project Knowledge
• Claude Code: Log entries written to filesystem
• Both use same format for compatibility