/status
Show knowledge stats for the current project
> /plugin marketplace add jamditis/claude-skills-journalismHow it fires
How this command gets triggered: by you, by Claude, or both.
- Fires itselfClaude auto-loads it when your prompt matches the work.
- You can call itInvoke it directly when you want it.
- Slash command
/status
Context preview
What this command does when you run it.
Show knowledge stats for the current project
Command definition
status.mddescription: Show knowledge stats for the current project
Display stats and status for the autocontext in the current project.
Precondition
Before displaying stats, the script checks if `.autocontext/` exists. If it doesn't, you'll see:
Autocontext not initialized for this project.
Run '/autocontext:init' to set up knowledge tracking.
Otherwise, the script reads:
- `.autocontext/lessons.json` — active and archived lessons
- `.autocontext/config.json` — project configuration
- `.autocontext/cache/` — any pending lessons
Status report
The script displays a clean, organized report with the following sections:
1. Lessons by category
Breakdown of all lessons (active and tombstoned) by category:
Lessons by category:
bug-fix: 12 active, 3 tombstoned
architecture: 8 active, 1 tombstoned
workflow: 15 active, 2 tombstoned
pattern: 6 active, 0 tombstoned
gotcha: 9 active, 1 tombstoned
integration: 4 active, 0 tombstoned
test-strategy: 5 active, 0 tombstoned
Total: 59 active lessons | 7 tombstoned
2. Confidence metrics
Average confidence across all active lessons:
Confidence metrics:
Average confidence: 0.72
High confidence (>= 0.8): 28 lessons
Medium confidence (0.5-0.8): 26 lessons
Low confidence (< 0.5): 5 lessons
3. Most-validated lessons
Top 5 lessons with the highest `validated_count` (most confirmed across sessions):
Most-validated lessons (by developer/session confirmations):
1. [Architecture] "Split domain logic from infrastructure" — validated 34 times
2. [Workflow] "Always run tests before commit" — validated 28 times
3. [Bug-fix] "Regex patterns need raw strings" — validated 22 times
4. [Pattern] "Use dependency injection for testing" — validated 18 times
5. [Test-strategy] "Write integration tests for API changes" — validated 16 times
4. Stalest lessons
Top 5 lessons with lowest confidence or oldest `last_validated` timestamp (need attention):
Stalest lessons (lowest confidence or oldest validation):
1. [Workflow] "Use Docker for local dev" — confidence 0.3, last validated 120 days ago
2. [Integration] "Slack API batch operations" — confidence 0.4, last validated 89 days ago
3. [Gotcha] "Python timezone handling edge case" — confidence 0.5, last validated 73 days ago
4. [Architecture] "Consider microservices for X" — confidence 0.4, last validated 95 days ago
5. [Pattern] "SQL query optimization with indexes" — confidence 0.5, last validated 67 days ago
5. Lessons by developer
Breakdown showing which developers discovered/contributed lessons:
Lessons by creator:
alice: 18 lessons
bob: 14 lessons
claude: 22 lessons
dave: 5 lessons
6. Recent lessons
Lessons added in the last 7 days:
Added in the last 7 days: 3 lessons
• [Bug-fix] "Array slice doesn't mutate original" — 2 days ago
• [Gotcha] "Event listener cleanup in unmount" — 4 days ago
• [Pattern] "Memoize expensive computations" — 6 days ago
If no lessons were added in the last 7 days, that's noted: `No lessons added in the last 7 days.`
7. Merge driver status
Shows whether cross-developer sharing is configured:
Merge driver status:
✓ Configured (git config merge.autocontext-union.driver found)
Or if not configured:
Merge driver status:
✗ Not configured (single-developer mode)
To enable, run: /autocontext:init
8. Pending lessons
Shows if there are lessons awaiting curation:
Pending lessons:
3 lessons in .autocontext/cache/pending-lessons.json (from ask_before_persist mode)
Run '/autocontext:review' to approve or reject.
Or if none:
Pending lessons:
None
Display format
The report uses:
- Clear section headers
- Indented, aligned data
- Counts and percentages where relevant
- Plain language (no technical jargon)
- Readable line breaks between sections
No interaction required
This skill is read-only. No AskUserQuestion is used. The report is informational, designed to help you understand the current state of your project's knowledge base.
Next steps
Based on the status output, you might:
- Run `/autocontext:review` if there are lessons with low confidence or pending approval
- Check stalest lessons to see if they're still valid or should be updated
- Share the status report with team members if using cross-developer sharing
Read more
description: Show knowledge stats for the current project
Display stats and status for the autocontext in the current project.
Precondition
Before displaying stats, the script checks if `.autocontext/` exists. If it doesn't, you'll see:
Autocontext not initialized for this project. Run '/autocontext:init' to set up knowledge tracking.
Otherwise, the script reads:
- `.autocontext/lessons.json` — active and archived lessons
- `.autocontext/config.json` — project configuration
- `.autocontext/cache/` — any pending lessons
Status report
The script displays a clean, organized report with the following sections:
1. Lessons by category
Breakdown of all lessons (active and tombstoned) by category:
Lessons by category: bug-fix: 12 active, 3 tombstoned architecture: 8 active, 1 tombstoned workflow: 15 active, 2 tombstoned pattern: 6 active, 0 tombstoned gotcha: 9 active, 1 tombstoned integration: 4 active, 0 tombstoned test-strategy: 5 active, 0 tombstoned Total: 59 active lessons | 7 tombstoned
2. Confidence metrics
Average confidence across all active lessons:
Confidence metrics: Average confidence: 0.72 High confidence (>= 0.8): 28 lessons Medium confidence (0.5-0.8): 26 lessons Low confidence (< 0.5): 5 lessons
3. Most-validated lessons
Top 5 lessons with the highest `validated_count` (most confirmed across sessions):
Most-validated lessons (by developer/session confirmations): 1. [Architecture] "Split domain logic from infrastructure" — validated 34 times 2. [Workflow] "Always run tests before commit" — validated 28 times 3. [Bug-fix] "Regex patterns need raw strings" — validated 22 times 4. [Pattern] "Use dependency injection for testing" — validated 18 times 5. [Test-strategy] "Write integration tests for API changes" — validated 16 times
4. Stalest lessons
Top 5 lessons with lowest confidence or oldest `last_validated` timestamp (need attention):
Stalest lessons (lowest confidence or oldest validation): 1. [Workflow] "Use Docker for local dev" — confidence 0.3, last validated 120 days ago 2. [Integration] "Slack API batch operations" — confidence 0.4, last validated 89 days ago 3. [Gotcha] "Python timezone handling edge case" — confidence 0.5, last validated 73 days ago 4. [Architecture] "Consider microservices for X" — confidence 0.4, last validated 95 days ago 5. [Pattern] "SQL query optimization with indexes" — confidence 0.5, last validated 67 days ago
5. Lessons by developer
Breakdown showing which developers discovered/contributed lessons:
Lessons by creator: alice: 18 lessons bob: 14 lessons claude: 22 lessons dave: 5 lessons
6. Recent lessons
Lessons added in the last 7 days:
Added in the last 7 days: 3 lessons • [Bug-fix] "Array slice doesn't mutate original" — 2 days ago • [Gotcha] "Event listener cleanup in unmount" — 4 days ago • [Pattern] "Memoize expensive computations" — 6 days ago
If no lessons were added in the last 7 days, that's noted: `No lessons added in the last 7 days.`
7. Merge driver status
Shows whether cross-developer sharing is configured:
Merge driver status: ✓ Configured (git config merge.autocontext-union.driver found)
Or if not configured:
Merge driver status: ✗ Not configured (single-developer mode) To enable, run: /autocontext:init
8. Pending lessons
Shows if there are lessons awaiting curation:
Pending lessons: 3 lessons in .autocontext/cache/pending-lessons.json (from ask_before_persist mode) Run '/autocontext:review' to approve or reject.
Or if none:
Pending lessons: None
Display format
The report uses:
- Clear section headers
- Indented, aligned data
- Counts and percentages where relevant
- Plain language (no technical jargon)
- Readable line breaks between sections
No interaction required
This skill is read-only. No AskUserQuestion is used. The report is informational, designed to help you understand the current state of your project's knowledge base.
Next steps
Based on the status output, you might:
- Run `/autocontext:review` if there are lessons with low confidence or pending approval
- Check stalest lessons to see if they're still valid or should be updated
- Share the status report with team members if using cross-developer sharing
A collection of Agent Skills for journalists, researchers, academics, media professionals, and communications practitioners. The same repository serves Claude Code and Codex while keeping Claude-only commands, agents, and hooks clearly labeled.
Repo: jamditis/claude-skills-journalism
Other commands on claude-skills-journalism.
- /evolve
Evolve skill files by integrating validated lessons from real usage
Open command - /init
Initialize .autocontext/ in the current project for knowledge persistence
Open command - /review
Interactively review and curate accumulated project lessons
Open command - /setup
First-run configuration for the autocontext plugin
Open command - /event
Create event materials (flyers, posters, signage) with your organization's branding
Open command - /newsletter
Create an HTML email newsletter with your organization's branding
Open command

