Conversation Recap

VerifiedCaution

Documents completed work sessions and updates knowledge base files. Automatically triggers when a plan is fully executed or all todos are marked complete, and can also be invoked manually with /recap. Analyzes the conversation, creates a session summary, and improves skills based on learnings.

Sby Skills Guide Bot
DocumentationIntermediate
6/2/2026
Claude Code
#conversation-recap#session-summary#knowledge-base#skill-improvement

Recommended for

Automated summary

Automatically documents completed work sessions by saving a structured summary and updating skill knowledge bases.

Strengths

  • Ensures knowledge capture after every meaningful task
  • Improves skills over time by integrating lessons learned
  • Creates auditable session logs for future reference
  • Reduces manual documentation effort through automation

Limitations

  • Only triggers on completed tasks, not quick Q&A sessions
  • May produce verbose summaries for trivial work
  • Relies on conversation logs, which may be compacted
When to use it

Use after finishing a plan, implementation, or research task to capture learnings and update knowledge bases.

When not to use it

Avoid for simple questions, minor fixes, or discussions with no lasting value.

Security analysis

Caution
Quality score88/100

The skill uses shell commands (ls, cat, jq) to access conversation logs within ~/.claude/projects/. These operations are scoped to well-known directories and serve a legitimate documentation purpose. No destructive or exfiltration commands are present, but the use of bash to parse files could expose sensitive conversation content if the skill is misapplied or logs are improperly handled.

No concerns found

How this analysis is made

Examples

Auto-recap on plan completion
All todos are marked complete. Run recap to document the session and update skills.
Manual recap with /recap
/recap

name: recap description: > Document the current conversation and update knowledge base files. USE THIS SKILL AUTOMATICALLY when a plan is executed to completion, all todos are marked complete, or meaningful work is finished. Can also be invoked manually with /recap command.

Conversation Recap

Summarize the session and improve skills based on feedback from the session.

When to Use

Auto-invoke this skill when:

  • A plan has been executed to completion (all planned steps done)
  • All todos in the todo list are marked complete
  • A significant implementation, analysis, or research task is finished

Skip if the conversation was just Q&A, quick fixes, or no lasting learnings.

Process

Step 0: Handle Compacted Conversations

If the conversation has been compacted (you see a summary at the start mentioning "continued from a previous conversation"), fetch the full conversation logs.

Log location: ~/.claude/projects/ (find the directory matching this workspace path)

# List project directories to find the right one
ls -d ~/.claude/projects/*/

# List recent conversation logs by size (larger = more content)
ls -laS ~/.claude/projects/<project-dir>/*.jsonl | head -10

# Extract user prompts from a conversation log
cat ~/.claude/projects/<project-dir>/<conversation-id>.jsonl | jq -c 'select(.type == "user") | .message.content[0].text' 2>/dev/null | grep -v "^null$" | head -30

Log entry types: user (prompts), assistant (responses), summary (compaction), progress (tools), system

Step 1: Analyze the Conversation

Review the conversation (and logs if compacted) to identify:

  1. Initial prompt: The user's original request
  2. Plan: The approved approach/plan
  3. Execution: What was done, step by step
  4. Result: Final outcome, deliverables, findings
  5. Errors/failures: Problems encountered and how they were solved
  6. Usage patterns: New skill usage patterns discovered
  7. Company context: New insights about the business

Step 2: Create Session Summary

Save to: workflows/yyyy-mm-dd-hh-mm-ss-workflow-title.md

Use the template from references/templates.md.

Filename guidelines:

  • Format: yyyy-mm-dd-hh-mm-ss-descriptive-title.md
  • Use lowercase with hyphens
  • Be descriptive but concise
  • Examples: 2026-01-25-14-30-00-email-revenue-analysis.md

Step 3: Improve Skills

Review the session for learnings that should be integrated into skills:

| Finding Type | Action | Location | |--------------|--------|----------| | Error/failure with solution | Add fix to skill's script or SKILL.md | .claude/skills/<skill>/ | | New usage pattern | Document in references | .claude/skills/<skill>/references/ | | New company context | Add to relevant department file | .claude/skills/company/references/ |

Department files: overview, team, goals, products, marketing, finance, operations, customer-service

Important: Ensure all skill revisions comply with skill-creator guidelines:

  • SKILL.md must stay under 500 lines
  • Split content into reference files when approaching this limit
  • Reference files from SKILL.md with clear descriptions of when to read them

Step 4: Summarize Changes

After documenting and updating, provide the user with:

  1. Path to the session summary file
  2. List of skill files that were updated
  3. Brief summary of what was captured

Guidelines

What to Include in Session Summary

  • Title with date: Descriptive title with date underneath
  • Initial prompt: Exact user request (quoted)
  • Plan: The approved approach
  • Execution summary: Step-by-step what was done
  • Result: Findings, deliverables, conclusions
  • Appendix: List of all skill revisions made

What to Add to Skills

For errors/failures:

  • Add solution to SKILL.md troubleshooting section, or
  • Fix the script directly in scripts/

For usage patterns:

  • Add to references/workflow-examples.md with example prompts and steps

For company context:

  • Add to the relevant department file in .claude/skills/company/references/

Don't

  • Create documentation for trivial conversations
  • Duplicate information across multiple files
  • Add verbose explanations - keep it concise
  • Exceed 500 lines in any SKILL.md file
  • Create files in .claude/plans/ (plans are temporary)

From the file to a first result

The download contains exactly the content displayed on this page. This copy may differ from the latest version in the source repository.

View the source and supporting files

This download contains SKILL.md only. Any scripts, references, templates and dependencies are not included. Get the required files from the source, preserving their directory structure, before continuing.

1. Download the file

Download SKILL.md

From the root of a test project, save the file as .skill-downloads/conversation-recap-mm5t8a2i/SKILL.md, or use the command below. This folder is for inspecting the skill before installation.

Command for macOS/Linux or WSL: requires a POSIX shell and curl. On Windows without WSL, use the download button and create the folder manually.

mkdir -p '.skill-downloads/conversation-recap-mm5t8a2i' && curl --fail --show-error --location 'https://skills-guides.com/api/skills/conversation-recap-mm5t8a2i/download' --output '.skill-downloads/conversation-recap-mm5t8a2i/SKILL.md.part' && mv '.skill-downloads/conversation-recap-mm5t8a2i/SKILL.md.part' '.skill-downloads/conversation-recap-mm5t8a2i/SKILL.md'

2. Check prerequisites

Read SKILL.md before using it. If it references scripts or external services, inspect them and prepare the required access in your test environment.

No recognized prerequisites section in this file. This does not mean it is self-contained: libraries, tools and access requirements may be mentioned elsewhere in the content or source repository.

The platforms on this page are catalog indications. The paths below follow tool documentation; they do not prove this skill has been run in those tools. No minimum tool version is confirmed for this listing.

3. Load the skill

After inspection, copy the complete folder into your project using one of the commands below, then open a new tool session. The command stops if a folder with the same name already exists: compare the files before replacing it.

Claude Code
test -f '.skill-downloads/conversation-recap-mm5t8a2i/SKILL.md' && test -s '.skill-downloads/conversation-recap-mm5t8a2i/SKILL.md' && mkdir -p '.claude/skills' && test ! -e '.claude/skills/recap' && cp -R '.skill-downloads/conversation-recap-mm5t8a2i' '.claude/skills/recap'

OpenSkills

OpenSkills must already be installed. The local install/read procedure is checked with OpenSkills 1.5.0, which requires Node.js ≥ 20.6.0. Catalog slugs are not resolved automatically.

Official OpenSkills installation
openskills install './.skill-downloads/conversation-recap-mm5t8a2i'

openskills read 'conversation-recap-mm5t8a2i'

Reading should display the installed file content and its base directory. It does not run the skill or install its dependencies.

For an agent that reads AGENTS.md, then sync the index. This command changes AGENTS.md: review the diff before using it.

openskills sync

Manual method or another tool: attach SKILL.md to a conversation, or open it in an agent that can read files. Include any required supporting resources. No universal folder is recognized by every tool.

4. Try a first task

Use a short, synthetic input: an API endpoint, a small document or an interface brief. In a new conversation, attach the file or give the agent its path, then adapt this request:

Read .skill-downloads/conversation-recap-mm5t8a2i/SKILL.md and apply its instructions to this example: [describe a test input]. Identify any missing files, dependencies or access. Produce the result requested by the skill without inventing data or running unreviewed scripts.

Expected result: output matching the skill instructions, or a precise list of missing prerequisites. A read confirmation alone does not validate a first use. If an external resource is unavailable, that path remains untested.

Related skills