New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Claude Session Observe

by danielwpz in danielwpz/pokoclaw

A skill your agent uses to inspect Claude Code session history under ~/.claude, recover what sessions exist, and manually determine what a session was doing from transcript evidence.

Apache-2.0Auto-check passed

Install Claude Session Observe

skills CLI
$ npx skills add danielwpz/pokoclaw --skill claude-session-observe -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install danielwpz/pokoclaw claude-session-observe --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/danielwpz/pokoclaw.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/claude-session-observe .claude/skills/claude-session-observe && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
claude-session-observe
GitHub stars
142
Token cost
~1.8k tokens
SKILL.md length
912 words
Files
8 (incl. scripts, references)
Skills in repo
4
Repo updated
First seen
Licence
Apache-2.0

At a glance

A skill your agent uses to inspect Claude Code session history under ~/.claude, recover what sessions exist, and manually determine what a session was doing from transcript evidence.

  • Works in 4 steps: User asks: what Claude Code sessions… → User asks: what was this session doing? → User asks: continue the latest Claude… → …
  • Inspect Claude Code session history under ~/.claude
  • SKILL.md covers Core principle, Channels, Required first reads and Default posture, plus 5 more sections
  • Runs Python scripts from its folder; calls git

What it does

Claude Session Observe is an agent skill from danielwpz/pokoclaw. Use this skill to inspect Claude Code session history under ~/.claude, recover what sessions exist, and manually determine what a session was doing from transcript evidence.

Its SKILL.md is about 1.8k tokens, which your agent loads only when the skill is triggered. The skill folder holds 9 other files, including scripts and reference files (for example `references/activity-heuristics.md`, `references/continuation-handoff.md` and `references/layout-overview.md`).

The repository describes itself as: The claw that actually ships 🚀. The licence is Apache-2.0.

When your agent uses it

  • Inspect Claude Code session history under ~/.claude
  • Recover what sessions exist
  • Manually determine what a session was doing from transcript evidence

Example prompts

  • “/claude-session-observe”

Requirements

  • Python 3

Workflow steps

4 steps, taken from the step headings in SKILL.md.

  1. User asks: what Claude Code sessions exist for this project?
  2. User asks: what was this session doing?
  3. User asks: continue the latest Claude Code session / finish what it did not finish
  4. User asks: check whether Claude's work actually landed

What it can do on your machine

Read from SKILL.md and the folder at commit 6cfb6c1. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 2 files in scripts/ (Python), which the agent can run.

    Shell commands in SKILL.md call:

    • git

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Claude Session Observe loads about 1.8k tokens when it runs, and up to ~6k if it reads all its reference files. Until then it costs about 49 tokens; SKILL.md has 912 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~49
When it runs · the whole SKILL.md, loaded when a task matches
~1.8k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~6k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from danielwpz/pokoclaw at commit 6cfb6c1, republished under its Apache-2.0 licence (© danielwpz). 912 words, ~1,782 tokens.

Download SKILL.mdSave it as .claude/skills/claude-session-observe/SKILL.md (or your agent's skills folder). This skill also uses 7 other files; get the full folder from GitHub.
name
claude-session-observe
description
Use this skill to inspect Claude Code session history under ~/.claude, recover what sessions exist, and manually determine what a session was doing from transcript evidence.
skillKey
pokoclaw/claude-session-observe

Claude Session Observe

Use this skill to inspect Claude Code's local session artifacts under ~/.claude.

The goal of this skill is not to teach you to trust a magic script. The goal is to teach you how to:

  • find which Claude Code sessions exist for a project
  • inspect what happened in a session
  • understand what a session was doing from transcript evidence
  • check whether work was completed, interrupted, delegated, or still needs follow-up
  • compare the session story with the current repo state when the user wants to continue or verify work

Core principle

There are three layers here:

  1. Artifacts
    • ~/.claude transcript and sidecar files
    • current repo/worktree state for the recovered cwd
  2. Query helpers
    • optional scripts under scripts/
    • ordinary file reads, ls, grep, tail-style inspection
  3. Your judgment
    • "what sessions exist"
    • "what this session was doing"
    • "what was completed"
    • "what should happen next"

The scripts are only layer 2. They are read-only retrieval helpers, similar to ls, tail, or select ... from ... limit .... You decide the final answer.

If the scripts do not exist, you should still be able to use this skill. The real contract of this skill is the investigation workflow.

Channels

  • Active session map: ~/.claude/sessions/*.json
  • Project session transcripts: ~/.claude/projects/<encoded-project>/<session-id>.jsonl
  • Session sidecar directory: ~/.claude/projects/<encoded-project>/<session-id>/
  • Subagent transcripts: subagents/*.jsonl
  • Subagent metadata: subagents/*.meta.json
  • Tool output captures: tool-results/*.txt
  • Global history: ~/.claude/history.jsonl
  • Current repo/worktree for the recovered cwd

Choose channels based on the question. Do not force every question through the same path.

Required first reads

  • If the task is about ~/.claude layout, path encoding, or where files live:
    • read references/layout-overview.md
  • If the task is about finding which session to inspect:
    • read references/session-discovery-recipes.md
  • If the task is about how to read transcript JSONL records:
    • read references/transcript-records.md
  • If the task is about interpreting recent work or takeover / continuation:
    • read references/activity-heuristics.md
    • then read references/continuation-handoff.md

Default posture

Start manually unless a helper script will obviously save repetitive work.

In many cases you do not need Python at all. The job can often be done with:

  • path resolution
  • directory listing
  • reading one or more transcript files
  • checking the tail of the transcript
  • checking the current repo state

Use scripts only when they make repetitive retrieval faster or safer.

Core workflows

1. User asks: what Claude Code sessions exist for this project?
  1. Resolve the project to an absolute path.
  2. Map it to the encoded project directory under ~/.claude/projects/.
  3. Check ~/.claude/sessions/*.json for active-session matches by cwd.
  4. List the project's *.jsonl files.
  5. Rank candidates by evidence:
    • active-session match
    • transcript timestamps
    • nearby progress / last-prompt
    • file mtime only as fallback
  6. Present the candidate sessions and your ranking evidence.
2. User asks: what was this session doing?
  1. Read the transcript tail.
  2. Inspect the latest useful evidence:
    • last-prompt
    • recent user records
    • recent assistant records
    • recent progress records
  3. If the parent transcript suggests delegated work, inspect subagents/.
  4. If a tool result is needed, inspect tool-results/.
  5. Then you explain:
    • what the user wanted
    • what Claude already did
    • whether the flow looks finished, interrupted, or still in flight

Do not ask the scripts to produce that explanation for you.

Show full SKILL.md (399 more words)Show less
3. User asks: continue the latest Claude Code session / finish what it did not finish
  1. Choose the best candidate session for the target project.
  2. Recover the session story from transcript evidence.
  3. Then inspect the current repo/worktree for that session's cwd.
  4. Compare transcript evidence with disk state.
  5. Answer with a continuation brief:
    • recovered objective
    • artifacts touched
    • completed so far
    • current repo/worktree state
    • what remains
    • immediate next action

Important: for continuation, transcript evidence alone is not enough.

4. User asks: check whether Claude's work actually landed
  1. Recover the claimed work from the transcript.
  2. Check the relevant repo directly:
    • git status --short --branch
    • git diff --stat
    • relevant file diffs or file reads
    • recent commits if needed
  3. Compare the claim with the actual repo state.
  4. Report agreement or mismatch explicitly.

Working rules

  • Prefer direct evidence over guesswork.
  • Prefer project-local transcript evidence over global history for project continuation.
  • Scripts are optional helpers, not answer generators.
  • If you use a script, treat its output as evidence to inspect.
  • Do not let a script decide the goal, stopping point, completion state, or next step.
  • Inspect progress before assuming the latest assistant text tells the whole story.
  • Inspect sidecar artifacts only when the transcript suggests they matter or when more detail is needed.
  • If the user wants continuation or verification, compare the session story with the current repo/worktree state.
  • If evidence is ambiguous, present ranked interpretations instead of pretending certainty.
  • If the trailing transcript line is malformed, ignore that line and continue with earlier valid records.

Optional helper scripts

These are convenience tools only.

  • scripts/find_project_sessions.py
    • read-only helper to list project-local candidate sessions
  • scripts/read_session_records.py
    • read-only helper to slice one session transcript and optionally list sidecar artifacts

Use them the same way you would use ls, tail, or a small select ... from ... limit ... query.

Available references

  • references/layout-overview.md
  • references/session-discovery-recipes.md
  • references/transcript-records.md
  • references/activity-heuristics.md
  • references/continuation-handoff.md

Anti-patterns

  • Do not treat helper scripts as semantic summarizers.
  • Do not expect a Python script to tell you what a session "means".
  • Do not stop at transcript reading when the user asked you to continue or verify work.
  • Do not ignore the current repo state for the recovered cwd.
  • Do not rely only on file mtime when transcript timestamps exist.
  • Do not assume tool results are top-level transcript records.
  • Do not assume interruptions have a dedicated top-level schema.
  • Do not require the scripts when direct file inspection is enough.

© danielwpz, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 7 other files (scripts, references) in skills/claude-session-observe of danielwpz/pokoclaw.

  • SKILL.md
  • references/activity-heuristics.md
  • references/continuation-handoff.md
  • references/layout-overview.md
  • references/session-discovery-recipes.md
  • references/transcript-records.md
  • scripts/find_project_sessions.py
  • scripts/read_session_records.py

Open the folder on GitHubat commit 6cfb6c1

Compare with similar skills

Claude Session Observe next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Claude Session Observe compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Claude Session Observe this skilldanielwpz/pokoclaw142—~1.8kAutomated safety check: PassApache-2.0
Langsmith ObservabilityOrchestra-Research/AI-Research-SKILLs13k3 repos~2.4kAutomated safety check: PassMIT
ObservabilityBuilderIO/agent-native7.1k—~7kAutomated safety check: PassNone
Frontend Observabilitysickn33/agentic-awesome-skills47k1 repos~5.1kAutomated safety check: PassMIT
Ebpf Observabilitysickn33/agentic-awesome-skills47k2 repos~3.3kAutomated safety check: NotesMIT
Python Observabilitywshobson/agents40k—~1.8kAutomated safety check: PassMIT

Similar skills

  • Langsmith Observability

    Orchestra-Research/AI-Research-SKILLs

    LLM observability platform for tracing, evaluation, and monitoring.

    13k GitHub starsUsed in 3 repos~2.4k tokens
    AI & LLM EngineeringAuto-check passed

  • Observability

    BuilderIO/agent-native

    Agent observability, evals, feedback, and experiments. An agent skill from BuilderIO/agent-native.

    7.1k GitHub stars~7k tokensUpdated today
    DevOps & CloudAuto-check passed

  • Frontend Observability

    sickn33/agentic-awesome-skills

    A portable, framework-agnostic field-side observability system for any React or React Native app.

    47k GitHub starsUsed in 1 repo~5.1k tokens
    DevOps & CloudAuto-check passed

  • Ebpf Observability

    sickn33/agentic-awesome-skills

    Use eBPF for deep kernel-level observability — trace syscalls, network flows, and application behavior without code changes using Cilium, Tetragon, and bpftrace.

    47k GitHub starsUsed in 2 repos~3.3k tokens
    DevOps & CloudAuto-check: notes

  • Python Observability

    wshobson/agents

    Python observability patterns including structured logging, metrics, and distributed tracing.

    40k GitHub stars~1.8k tokensUpdated 3 days ago
    DevOps & CloudAuto-check passed

  • Observability Designer

    alirezarezvani/claude-skills

    Design production-ready observability strategies combining metrics, logs, and traces.

    28k GitHub stars~3.5k tokensUpdated 1 mo ago
    DevOps & CloudAuto-check passed

More from danielwpz/pokoclaw

  • A2ui Author

    danielwpz/pokoclaw

    Generate, validate, and interpret static A2UI v0.8 JSON for interactive UI surfaces.

    142 GitHub stars~330 tokensUpdated 19 days ago
    Auto-check passed

  • Skill Installer

    danielwpz/pokoclaw

    Install third-party skills for Pokoclaw when the user asks an agent to install a skill, pastes a skill-store prompt, or pastes a skill-store CLI command.

    142 GitHub stars~1.9k tokensUpdated 19 days ago
    Auto-check passed

  • System Observe

    danielwpz/pokoclaw

    Use this skill before calling querysystemdb or when diagnosing pokoclaw.

    142 GitHub stars~1.5k tokensUpdated 19 days ago
    Auto-check passed

Questions about Claude Session Observe

What does Claude Session Observe do?

A skill your agent uses to inspect Claude Code session history under ~/.claude, recover what sessions exist, and manually determine what a session was doing from transcript evidence. Claude Session Observe is an agent skill from danielwpz/pokoclaw.claude, recover what sessions exist, and manually determine what a session was doing from transcript evidence.

When should I use Claude Session Observe?

Claude Session Observe fits situations like: inspect Claude Code session history under ~/.claude; recover what sessions exist; manually determine what a session was doing from transcript evidence.

How do I install Claude Session Observe in Claude Code?

Run `npx skills add danielwpz/pokoclaw --skill claude-session-observe -a claude-code`. Or copy the skill folder (skills/claude-session-observe in danielwpz/pokoclaw) into .claude/skills/claude-session-observe in your project. Claude Code loads it when a task matches its description.

How do I install Claude Session Observe in Codex?

Run `npx skills add danielwpz/pokoclaw --skill claude-session-observe -a codex`. Or copy the skill folder (skills/claude-session-observe in danielwpz/pokoclaw) into .agents/skills/claude-session-observe in your project. Codex loads it when a task matches its description.

Can I use Claude Session Observe in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add danielwpz/pokoclaw --skill claude-session-observe -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/claude-session-observe, .gemini/skills/claude-session-observe, .github/skills/claude-session-observe and .opencode/skills/claude-session-observe in your project.

What does Claude Session Observe need to run?

Going by SKILL.md and its folder, Claude Session Observe needs Python for the scripts in its folder and the command-line tools its instructions call (git). Our summary lists: Python 3.

Does Claude Session Observe access the network?

SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Claude Session Observe safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Claude Session Observe use?

Claude Session Observe is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Claude Session Observe use?

About 1.8k tokens (SKILL.md is roughly 7.1k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 4.2k tokens, read only when the agent opens those files.

What are the alternatives to Claude Session Observe?

Skills that share tags, products or a category with Claude Session Observe: Langsmith Observability (Orchestra-Research/AI-Research-SKILLs, 13k stars), Observability (BuilderIO/agent-native, 7.1k stars), Frontend Observability (sickn33/agentic-awesome-skills, 47k stars) and Ebpf Observability (sickn33/agentic-awesome-skills, 47k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Claude Session Observe?

danielwpz (a GitHub user) maintains it in danielwpz/pokoclaw, which has 142 GitHub stars. The repository holds 4 skills in this directory. The repository was last updated on September 19, 2026.

Source: danielwpz/pokoclaw on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.