Install
$ agentstack add skill-brettrey-claude-academic-skills-session-management ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Session Management
LLM sessions are stateless. When a session ends, everything learned -- decisions, preferences, project state, open questions -- evaporates. This skill creates a persistence layer using two files so the next session can pick up where the last one left off.
Persistence Files
STATUS.md -- Human-readable project status. Dated entries with bullet points. This is the primary file for context recovery.
session-log.jsonl -- Machine-readable session log. One JSON line per session. Useful for scanning recent activity across many sessions.
Both files are created automatically if they don't exist.
Shutdown Protocol
Use when the user says "shutdown", "wrap up", "done for now", or is ending a session.
Step 1: Session Summary
Review what was accomplished this session. Identify:
- Decisions made that should persist (design choices, rejected approaches, user preferences)
- Context learned (constraints discovered, patterns noted, things that didn't work)
- State changes (files created or modified, tasks completed, blockers found)
- Open questions or next steps that the next session should start with
Step 2: Update STATUS.md
Add a dated entry to STATUS.md in the project root. Create the file if it doesn't exist.
Format:
### YYYY-MM-DD
- [Decision or state change]
- [What was accomplished]
- [Open questions or next steps]
Most recent entries go at the top, below the heading.
Step 3: Append to Session Log
Append one line to session-log.jsonl in the project root. Create the file if it doesn't exist.
{"timestamp": "YYYY-MM-DDTHH:MM:SSZ", "project": "project-name", "summary": "one-line summary of session"}
The summary should be a single sentence capturing the session's main accomplishment.
Step 4: Context File Interview
Ask the user one question before closing:
> "Anything from this session that should be added to a context file (CLAUDE.md or similar)?"
If yes, make the updates. If no, proceed. Do not skip this step. It is the highest-value action for preventing context loss -- the user often knows things that aren't captured in the summary.
Step 5: Confirm Exit
> "Session summary saved to STATUS.md and session-log.jsonl. Ready to exit?"
Startup Protocol
Use when a new session begins, or the user says "pick up where I left off", "what did we do last time", or similar.
Step 1: Read STATUS.md
Find STATUS.md in the project root. Read the most recent dated entry. Note decisions, state changes, and open questions.
Step 2: Read Session Log
Read the last 5-10 entries of session-log.jsonl. If multiple sessions have happened since the user last worked on this project, summarize the arc (not just the last session).
Step 3: Check for Open Items
If the most recent STATUS.md entry mentions blockers, open questions, or explicit next steps, collect them.
Step 4: Brief the User
> "Last session (YYYY-MM-DD): [summary]. Open items: [list or 'none']. Ready to continue?"
Keep the briefing short. The user can ask for details if needed.
Long Session Recovery
In long sessions, early context fades as the conversation grows. If the user seems disoriented, asks "where were we", or has been working for a long time without a clear thread, re-read STATUS.md and session-log.jsonl to recover the session's starting state and trajectory. Summarize what's been done and what remains, then resume.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: BrettRey
- Source: BrettRey/claude-academic-skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.