agentsclimarketplace

Session resume

Skill TheophilusChinomona/idev/skills/session-resume

Cross-session context persistence via .claude/idev/session-resume/last-session.json. Use when resuming work in a new session, or to save working context at task completion or when the user asks to save.From its SKILL.md

Install
npx -y skills add TheophilusChinomona/idev --skill session-resume

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.5 KB, 812 tokens by cl100k_base, as published. Nobody here has run it

Session Resume Skill

Purpose

Persist working context across sessions so Claude can resume exactly where it left off without re-discovering files, re-reading code, or re-learning what was being worked on.

Activation

  • Save: At task completion, or when the user asks to save context
  • Load: At the start of a new session (before any task work)

(A PreCompact hook could automate saving before context compaction, but none is wired by default — don't rely on it.)


How It Works

A lightweight session snapshot is saved to: .claude/idev/session-resume/last-session.json

This file is loaded at session start, giving Claude instant awareness of:

  • What feature/area was being worked on
  • Which files were recently read/modified
  • What the last task was
  • Any pending issues or blockers

Phase 1: Save Session State

At task completion (or when the user asks to save context), save:

{
  "savedAt": "YYYY-MM-DD HH:mm",
  "lastTask": {
    "description": "Brief description of what was being done",
    "status": "completed|in-progress|blocked",
    "blockedBy": "Description of blocker (if blocked)"
  },
  "activeFeature": "FeatureName (e.g., Orders)",
  "activeLayer": "frontend|backend|full-stack",
  "recentFiles": {
    "modified": [
      "relative/path/to/file1.ts",
      "relative/path/to/file2.cs"
    ],
    "read": [
      "relative/path/to/file3.ts"
    ]
  },
  "openIssues": [
    {
      "description": "Reject button not hitting API - needs browser console check",
      "file": "relative/path/to/file.tsx",
      "severity": "medium"
    }
  ],
  "workingBranch": "feature/branch-name",
  "notes": "Any important context that would be lost on session restart"
}

Phase 2: Load Session State

At session start:

1. Read last-session.json (~30 lines, roughly 300 tokens)
2. If lastTask.status == "in-progress":
   → Inform user: "Last session was working on: [description]"
   → Offer to continue or start fresh
3. If openIssues exist:
   → Mention pending issues briefly
4. Use activeFeature + activeLayer to pre-load the right context
5. Use recentFiles.modified to know which files were changed

Phase 3: When to Save

Save the session state when ANY of these occur:

  1. User explicitly says "save session" or "save context"
  2. A task is completed (update lastTask.status)
  3. User switches to a different feature/task
  4. A blocker is encountered (update openIssues)

Save Rules

  • Keep recentFiles lists to max 10 entries each (most recent first)
  • Keep openIssues to max 5 entries
  • Overwrite previous session file (only latest session matters)
  • Do NOT save file contents — only paths
  • Do NOT save sensitive data (tokens, passwords, keys)
  • Total file size should stay under 50 lines

Integration with Other Skills

Session start:
  1. Load last-session.json (this skill)
  2. Load .claude/idev/smart-context/index.json (smart context skill)
  3. If activeFeature is set → grep file-index for that feature's files
  4. Ready to work in ~3 reads

Anti-Patterns

  1. Do NOT save entire file contents in the session state
  2. Do NOT save more than 10 recent files
  3. Do NOT load the session state AND re-explore the codebase
  4. Do NOT ignore the session state when it exists

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most hr recruiting skills give in 812 tokens

Counted across 356 of the 357 authors here whose files we hold, read 2026-08-07

  • Quantify achievements with specific metricsin 14 of 356, across 6 files
  • Keep the resume under two pagesin 14 of 356, across 6 files
  • Request the full job description if not providedin 12 of 356, across 4 files
  • Extract keywords and prioritize job requirementsin 12 of 356, across 4 files
  • Stop and ask for clarification if required inputs are missingin 12 of 356, across 5 files
  • Map candidate experience to job requirementsin 11 of 356, across 3 files
  • Ask if the user wants adjustmentsin 11 of 356, across 3 files
  • Provide strengths and gap analysis after the resumein 10 of 356, across 2 files
  • Request candidate background details if not providedin 10 of 356, across 2 files
  • Format experience bullets as action verb plus resultin 10 of 356, across 2 files
  • Ask for missing inputs before startingin 10 of 356, across 9 files
  • Use exact job description terminologyin 9 of 356, across 1 file

Said here and by no other author read

  • save session state at task completion
  • save session state on user request
  • save session state when encountering a blocker
  • load session state at new session start
  • limit recent file lists to ten entries
  • limit open issues to five entries

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.