Session SaveCAUTION
The most comprehensive Claude Code guide: agentic workflows, hooks, skills, MCP servers, quizzes, and production-ready templates. 430K+ lines.
Overview
The most comprehensive Claude Code guide: agentic workflows, hooks, skills, MCP servers, quizzes, and production-ready templates. 430K+ lines.
d90170da4369OBSERVED · 2026-10-07What it tells the agent
The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.
--- name: session-save description: Save the current session state (decisions, modified files, current status, and next steps) to a handoff file for later resume. effort: low disable-model-invocation: true --- # /session-save Capture the current session into a structured handoff file so you (or another instance) can resume with full context. Creates a timestamped Markdown file in `.claude/sessions/`. ## When to use - Before ending a session that isn't complete - Before switching to a different task - Before context reaches 75%+ (to preserve the important parts) - At natural milestones in a long task (after each phase) - To hand off to a second Claude instance ## Instructions Produce a handoff document with the following structure: --- ## Session handoff: [TIMESTAMP] ### What was being done [One paragraph: the goal, the approach, where things stand right now] ### Files modified this session [List every file that was created, edited, or deleted, with the nature of the change] ``` path/to/file.ts - [what changed and why] path/to/other.ts - [what changed and why] ``` ### Key decisions made [Architectural choices, tradeoffs accepted, approaches rejected and why] - **Decision**: [What was decided] - **Rationale**: [Why] - **Alternatives rejected**: [What else was considered] ### Current status [Where things are right now: what's working, what's broken, what's in-progress] - Working: [...] - In-progress: [...] - Known issues: [...] ### Next steps (ordered) [The exact next actions to take to continue, specific enough that a fresh context can pick up without re-reading everything] 1. [First action]: `path/to/file.ts` - [what to do] 2. [Second action]: [...] 3. [...] ### Context to reload [Files that must be read to resume with full understanding, keep this list short] - `path/to/key-file.ts`: [why it matters] - `CLAUDE.md`: project rules ### Blockers / open questions [Anything unresolved that needs a decision or external input before proceeding] - [ ] [Question or blocker]: [who/what can resolve it] --- ## Implementation Save the handoff to `.claude/sessions/handoff-[YYYY-MM-DD-HHMM].md`. Then output the file path so the user knows where to find it. ## Resume pattern To resume from a handoff: ``` /session-resume .claude/sessions/handoff-YYYY-MM-DD-HHMM.md ``` Or manually: read the handoff file, then read the files listed in "Context to Reload" before continuing. ## Examples ### Example 1: Mid-Feature save ``` /session-save ``` Claude captures: - Auth middleware was being refactored (files: `src/middleware/auth.ts`, `src/middleware/jwt.ts`) - Decision: switched from session tokens to JWTs (rationale: scales better across services) - Status: JWT validation working, refresh logic in-progress - Next: implement `refreshToken()` in `src/services/auth.service.ts`, then update tests ### Example 2: Context-Pressure save When context hits 70%, run `/session-save` before `/compact` to preserve decision context that compaction might lose. ## Notes - Keep "Context to Reload" to 5 files max. The goal is a fast resume, not a full re-read. - "Next Steps" should be specific enough that a cold-start Claude can execute step 1 without asking questions - Don't save tool output or code snippets in the handoff; reference file paths instead. --- **See also**: - [Session Teleportation](../../../guide/ultimate-guide.md#916-session-teleportation): broader session management patterns - [Instinct-Based Learning](../../../guide/ultimate-guide.md#924-instinct-based-continuous-learning): what to extract from sessions before closing
Trust audit
CAUTIONgrade B · trust 89/100 Install with care. The audit found things worth knowing before you trust its output.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | WARN |
| L1 | Static analysis of the code | NA |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (2)
whitepapers/recap-cards/en/_extensions
whitepapers/recap-cards/fr/_extensions
Gates applied: no_behavioural_pass.
d90170da4369full audit observations/trust-audit/skill/florianbruniaux__session-save.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-07 | d90170da4369 | CAUTION | B | 89 | first audit |
Questions
What does the Session Save skill do?
The most comprehensive Claude Code guide: agentic workflows, hooks, skills, MCP servers, quizzes, and production-ready templates. 430K+ lines.
Is Session Save safe to install?
With care. The audit graded it B (89/100) and found 2 things worth knowing before you trust this skill, listed below with the exact line each was found on.
What can Session Save access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
How current is this page?
The grade is for one exact copy of the source (d90170da4369), read on 2026-10-07. The repository is watched, and a new audit runs when it changes — this is the first audit.