session-start-routine
Pre-flight checks at session start — load context, check prior state, validate env, check for in-flight work from other terminals
Best use case
session-start-routine is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Pre-flight checks at session start — load context, check prior state, validate env, check for in-flight work from other terminals
Teams using session-start-routine should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/session-start-routine/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How session-start-routine Compares
| Feature / Agent | session-start-routine | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Pre-flight checks at session start — load context, check prior state, validate env, check for in-flight work from other terminals
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Session Start Routine
Pre-flight checklist for every new session. Run these checks before beginning work.
## Checklist
### 1. Load context
- Load memory context files from the repo memory directory
- Check `MEMORY.md` for active project state and recent feedback
### 2. Check prior session state
- Read today's session signals: `.Codex/state/session-signals/YYYY-MM-DD.jsonl`
- Look for sessions that ended mid-task (incomplete commits, unreleased `wip:` labels)
- Check `git status` for uncommitted work from prior sessions
### 3. Check for in-flight work
- Scan for wip labels on GitHub issues (gh issue list filtered by wip label)
- Check other terminals: `ps aux | grep Codex | grep -v grep`
- Check `/tmp/.Codex-wip-*` markers if present
### 4. Validate environment
- Verify tools: `uv --version`, `gh auth status`, `git status`
- Check disk space: `df -h /mnt/local-analysis`
- Verify governance hooks are registered in settings
### 5. Check governance limits
- Review tool-call counter: `.Codex/state/session-governor/tool-call-count`
- Confirm counter is below 200 ceiling (or reset if new day)
## When to use
- Start of every interactive session
- After a context reset or conversation compression
- When resuming work after a break
- When the user asks for a terse live "status update" from mobile/Telegram
## Mobile status-update mode
When the user asks only "Status update" (especially from Telegram), produce a compact evidence-backed snapshot rather than a broad narrative.
1. Gather live state before summarizing:
- `git status --short --branch` in the active repo.
- Active/background agent processes or Hermes process-manager state.
- `hermes cron`/cronjob list for active, paused, failed, and next scheduled jobs.
- GitHub counts for `wip`, `status:plan-review`, `status:plan-approved`, and recent closed issues.
- Most recent relevant cron output files when scheduled jobs produced user-facing findings.
2. Separate counts clearly:
- live label counts are not the same as artifact-audited approval readiness;
- closed/recent results are not the same as currently running lanes;
- paused autofeed monitors are not active throughput.
3. For Telegram/mobile delivery:
- avoid pipe tables; use numbered bullets and labeled key/value lines;
- link GitHub issues as `[#{number}](url)`;
- keep to the shortest useful update, with risks and next best action last.
4. Never clean/reset/stage dirty control-plane files during a status-only request; report dirty state and recommend a narrow reconciliation pass if needed.Related Skills
provider-session-ecosystem-audit-and-exporters
Build and maintain cross-provider session-log audits for Codex, Codex, Hermes, and Gemini, including exporter design, normalization, and behavioral verification.
tax-filing-session-setup-with-github-tracking
Structured workflow for preparing and tracking a tax filing session using prepared documents, task checklist, and GitHub issue cross-referencing
tax-filing-session-setup-with-github-traceability
Structured workflow for setting up a multi-file tax filing session with GitHub issue tracking and prepared-file validation
handle-session-overlay-blocking
Technique for dismissing overlay dialogs that freeze rendering and block form interaction in web automation
handle-freetaxusa-session-timeouts
Recover from FreeTaxUSA session timeout dialogs blocking form submission and navigation
extract-skills-from-hermes-sessions
Automatically analyze Codex session transcripts to identify and extract reusable skills using LLM analysis via OpenRouter
extract-skills-from-Codex-sessions
Automatically extract reusable skills from Codex session transcripts using LLM analysis and wire them into a Stop hook
session-start-dirty-state-triage-with-background-agents
Distinguish real implementation dirt from generated session-state churn when resuming a repo with active overnight/background agents.
session-corpus-audit
Analyze session quality trends — identify high-churn patterns, report waste, flag sessions exceeding 500 tool calls
provider-session-learning-transfer
Refresh provider session audit, identify post-audit/unassessed sessions, extract actionable learnings, and transfer them into repo notes and GitHub issues before a follow-up implementation session.
provider-session-ecosystem-audit
Audit Codex/Codex/Hermes/Gemini session logs, normalize provider-specific quirks, and wire recurring exports/reporting for ongoing ecosystem health checks.
provider-session-quota-operations
Class-level provider/session operations for Codex, Codex, Gemini, Hermes, quotas, audit exporters, readiness dispatch, and utilization scorecards.