open

name	open
description	Companion skill for the /open workflow. Automates workspace diagnostic steps when starting a new session on a project. Collects active session status, project contracts, active task metrics, compact memory presence, roadmap phases, pending brainstorms, and git status.
version	1.9.5

Open Skill

Purpose: Automate project opening checks and gather developer context. Trigger: Active during the /open workflow execution, or when starting/resuming a project session.

Workflow

This skill is executed during the initialization of a project session to check for configuration, Git status, and task files.

Step 1: Execute Project Diagnostics

Run the script scripts/run_open.sh with the target project name as the first argument.

bash .agents/skills/open/scripts/run_open.sh [project-name]

Step 2: Parse and Report

Parse the output of the script to construct the session opening report. Identify any non-zero exit codes. Note that exit codes of 1 from grep operations are normal if no matching patterns exist.

Input / Output

Input

• Format: String (project name)
• Example:

pageel-crm

Output

• Format: Shell output stdout/stderr containing diagnostic details.
• Example:

=== DIAGNOSTICS FOR PROJECT: pageel-crm ===
Active session: Inactive
Latest sessions: Projects/pageel-crm/sessions/2026-06-03.md ...
Pending brainstorms: None
Compact memory: Present
Sprint-current lines: 12
Active backlog tasks: 0
Git status: Modified files ...
Pending Specifications: spec-2026-07-09-debug-observability-convention.md (⏳ Pending)
==========================================

Edge Cases

• Project directory not found: Script exits with code 1 and logs error.
• Missing project.md: Script warns and skips contract parsing.
• No git repository (repo/): Script skips git checks.

Scripts

Read and run scripts/run_open.sh when performing project diagnostics.

🧪 Test Mode (Sandbox Override)

Trigger: User includes "Test Mode" or explicitly asks to evaluate/test this skill.

When in Test Mode, follow these overrides:

1. No Live Edits: Do NOT modify files outside the sandbox directory.

2. Containment: Route ALL outputs into [PROJECT_ROOT]/sandbox/evals/open-[YYYY-MM-DD]/.

3. Execute Task: Run diagnostics for the mock project or target project and output the test report.

4. Generate Report: After completing the task, create test-report.md in the sandbox folder:

   # Test Report: open
   > Date: YYYY-MM-DD | Prompt: "[user's prompt]"
   
   ## Actions Taken
   - Run open diagnostics on the sandbox project.
   
   ## Skill Rules Invoked
   - Step 1: Run diagnostics script.
   
   ## Files Created
   - sandbox/evals/open-[date]/test-report.md
   
   ## Self-Assessment
   - Diagnosed successfully. Verified exit codes and output formatting.