agentleFS
fizzfizz/claude-skills/session-start/SKILL.md

session-start

Skill0 starsChanged 8 months agoMITView on GitHub

Use at the beginning of any session to orient — checks active branches across repos, reads all recent handoffs, queries Jira across all projects, checks GitHub PRs, and prevents stale-context errors

42 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use session-end, deploy-verify and 3 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • jasonmichaelbell78-creatorSkill

    session-end

    Also in 12 projects that use this

  • danielarif26Skill

    deploy-verify

    Verify a Vercel deployment actually went live and serves the correct production domain. Self-improving: reads its own run history before checking and logs the outcome after. Use when the user asks "did you deploy?", "is it live?", "verify the deploy", "check vercel", or after any Vercel deploy. Part of the Agentic OS (Level 1 backbone).

    Also in 3 projects that use this

  • Microsoft AzureSkill

    reflect

    Learning capture system that extracts HIGH/MED/LOW confidence patterns from conversations to prevent repeating mistakes. Use after user corrections ("no", "wrong"), praise ("perfect", "exactly"), or when discovering edge cases. Complements .squad/agents/{agent}/history.md and .squad/decisions.md.

    Also in 4 projects that use this

  • ibahgatSkill

    autopilot

    Run autonomous multi-stage execution loops until acceptance criteria pass or blockers are reached.

    Also in 3 projects that use this

  • MicrosoftSkill

    handoff

    Compact the current conversation into a handoff document for another agent to pick up.

    Also in 3 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. 0 are common, 14 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 35 edited versions we read keep these.

  • 14

    Distinctive

    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
  1. Distinctive:

    Start of any new session

    line 14
  2. Distinctive:

    If the most recent handoff mentions work in other repos, check those too:

    line 37
  3. Distinctive:

    Read ALL of them, not just the last one.

    line 51
  4. Distinctive:

    If you work across multiple Jira projects, query all of them:

    line 60
  5. Distinctive:

    If you use multiple Git hosting platforms (Bitbucket, GitLab, etc.), check those too:

    line 88
  6. Distinctive:

    If you own the CI/CD surface and other team members create their own workflows, detect changes and quality issues:

    line 96
  7. Distinctive:

    Keep it dense. This is orientation, not a report.

    line 160
  8. Distinctive:

    If the session involves project-specific work:

    line 164
  9. Distinctive:

    Do NOT start implementing anything until orientation is complete

    line 173
  10. Distinctive:

    Read ALL recent handoffs, not just the most recent — some users do 8 sessions/day

    line 174
  11. Distinctive:

    Check multiple repos if the handoff mentions cross-repo work

    line 176
  12. Distinctive:

    Check all Jira projects, not just the main one

    line 177
  13. Distinctive:

    Check all Git hosting platforms for PR state

    line 180
  14. Distinctive:

    Only checking one Jira project

    line 189

“Common” is measured: kept by at least 6 in 10 of the 35 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Session Start

Overview

Full orientation at session start. Prevents "loaded wrong plan," "wrong branch," "stale ticket," and "missed overnight activity" errors by checking actual state across all surfaces before doing anything.

When to Use

  • Start of any new session
  • Resuming after a break
  • User asks "where are we?" or "what's the status?"

Process

1. Git state

# Fetch and prune stale remote-tracking branches
git fetch --prune

# Current branch and recent commits (since last handoff, not arbitrary 5)
git branch --show-current
git log --oneline --since="yesterday"

# Any uncommitted work?
git status

# Stale local branches tracking deleted remotes (merged PRs that weren't cleaned up)
git branch -vv | grep ': gone]'

If the most recent handoff mentions work in other repos, check those too:

git -C /path/to/other/repo fetch --prune
git -C /path/to/other/repo log --oneline --since="yesterday"

2. Read recent handoffs

# ALL handoffs from today and yesterday — not just the most recent
ls -t docs/HANDOFF-$(date +%Y-%m-%d)-session-*.md docs/HANDOFF-$(date -v-1d +%Y-%m-%d)-session-*.md 2>/dev/null

Read ALL of them, not just the last one. Some users do up to 8 sessions per day. Each handoff has:

  • Completed items — verify they're actually done, don't re-list
  • In-progress items — verify branch/PR still exists
  • TODOs with @done tags — respect them, never re-list as pending
  • Harvest sections — scan for unharvested bonbons, blog material, reusable patterns
  • Reminders for next session — these are instructions, follow them

3. Check active work items (all projects)

If you work across multiple Jira projects, query all of them:

# Primary project
JIRA_PROJECT=PROJ jira ls -q "assignee = currentUser() AND status = 'In Progress' ORDER BY updated DESC"

# Secondary projects (add as many as needed)
JIRA_PROJECT=PROJ2 jira ls -q "project = PROJ2 AND assignee = currentUser() AND status = 'In Progress' ORDER BY updated DESC"

JIRA_PROJECT=PROJ3 jira ls -q "project = PROJ3 AND assignee = currentUser() AND status = 'In Progress' ORDER BY updated DESC"

# Recently updated tickets across ALL projects (last 24h) — catches status changes, new comments, new tickets
JIRA_PROJECT=PROJ jira ls -q "updated >= -1d ORDER BY updated DESC"

4. Check PRs across platforms

GitHub:

# Open PRs across your org's repos
gh pr list --repo your-org/your-repo --state open
gh pr list --repo your-org/your-api --state open
gh pr list --repo your-org/your-frontend --state open

# Recent workflow failures
gh run list --repo your-org/your-repo --limit 5 --json conclusion,workflowName --jq '.[] | select(.conclusion == "failure")'

If you use multiple Git hosting platforms (Bitbucket, GitLab, etc.), check those too:

# Use the appropriate CLI or API for each platform
# Check all repos with active development for open PRs

5. Check GHA workflow hygiene

If you own the CI/CD surface and other team members create their own workflows, detect changes and quality issues:

# New or modified workflow files in the last 7 days
git log --since="7 days ago" --name-only --diff-filter=ACMR -- '.github/workflows/*.yml' | sort -u

# For repos not cloned locally, compare HEAD against a week-old commit
# gh api "repos/OWNER/REPO/compare/OLDSHA...HEAD" --jq '.files[] | select(.filename | startswith(".github/workflows/"))'

# Recent workflow runs — look for sloppy patterns:
# - Same workflow run 5+ times in a row (trial-and-error without fixing the root cause)
# - Runs without meaningful titles (no context for future debugging)
# - Failed runs that were never investigated
gh run list --repo your-org/your-repo --limit 30 --json workflowName,conclusion,displayTitle,createdAt

Flag for review:

  • New workflow files you haven't reviewed
  • Workflow runs with empty or default titles (no displayTitle or just the commit message)
  • The same workflow failing 5+ times in a row (someone is spamming reruns instead of debugging)
  • Any workflow that bypasses plan-before-apply or skips tests

Also watch for infra drift — resources created via kubectl or console that aren't in terraform, and untagged infrastructure. This includes live patches applied during firefighting that never made it into the IaC module. If the last handoff mentions live patches, check whether they've been baked into terraform yet.

This is protagonist support for the shared CI/CD space — making sure the infrastructure stays legible and accountable.

6. Check pipeline state

For each in-progress work item, trace through the pipeline:

StageHow to check
Ticket createdJira query (step 3)
Branch existsgit branch -a | grep TICKET
PR opengh pr list --search TICKET or platform API
PR mergedgh pr list --state merged --search TICKET
Needs approvalCheck last Jira comment / Slack thread
DeployedCheck workflow runs

7. Summarize

Present a brief orientation:

Branch: feat/PROJ-481-copilotkit-sidecar
Stale branches: fix/PROJ-482-apiserver-stability (remote deleted)

Handoffs read: 2 from today, 1 from yesterday
  - Session 01: ETL seed job completed, service decoupled
  - Session 02: design investigation, synthesis posted

Active tickets (across projects):
  - PROJ-483 (EKS node auto-repair) — To Do
  - PROJ-488 (demo response epic) — To Do
  - PROJ2-434 (FinOps spike) — In Progress

Open PRs: none
Failed workflows: none

Unharvested from handoffs:
  - bonbon: "the format is not the signal"
  - blog idea: template skills vs behavior skills

Keep it dense. This is orientation, not a report.

8. Load project-specific skills and briefing

If the session involves project-specific work:

  1. Load any Jira-related skills — if your org uses Jira heavily, every session will touch tickets.
  2. Load any briefing/morning-paper skill — pulls Slack channels, cloud support cases, meeting transcripts, and semantic search into a full situational briefing.

The morning paper builds on the orientation from steps 1-7. Session-start provides the facts. Morning paper provides the narrative.

Rules

  • Do NOT start implementing anything until orientation is complete
  • Read ALL recent handoffs, not just the most recent — some users do 8 sessions/day
  • Respect @done tags — if a TODO has @done, it's done. Period.
  • Check multiple repos if the handoff mentions cross-repo work
  • Check all Jira projects, not just the main one
  • git fetch --prune before reporting branch state
  • Report stale local branches that track deleted remotes
  • Check all Git hosting platforms for PR state

Red Flags

  • Starting implementation without checking current branch
  • Ignoring older handoffs from the same day
  • Assuming ticket states without checking — statuses change between sessions
  • Working on a ticket that's already Done
  • Reporting on a branch that was merged and deleted upstream
  • Only checking one Jira project
  • Missing PR activity on secondary platforms
  • Re-listing items marked @done in handoff TODOs

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.