OpenSkillIndex← back to search
claude-code

Clawpage clawpage

Share AI agent conversations as public web pages. Use when the user wants to share a conversation externally, export conversation history for documentation, or publish a chat session to a public URL.

install
source · Clone the upstream repo
git clone https://github.com/imyelo/clawpage
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/imyelo/clawpage "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/clawpage" ~/.claude/skills/imyelo-clawpage-clawpage && rm -rf "$T"
manifest: skills/clawpage/SKILL.md
source content

clawpage

Share AI agent conversations as public web pages.

Supported Agents

AgentProfile
OpenClawreferences/platforms/openclaw.md
(unknown)references/platforms/unknown.md — generic skill-based fallback
(new platform)Add a file following references/platforms/TEMPLATE.md

Core Workflow

1. Setup Check

  • Detect agent type; load project dir + site URL using the agent profile
  • If project not configured locally, ask the user:

2. Locate Session

  • List sessions using agent profile discovery
  • Default to the most recent (current) session; offer to pick any historical session by ID or keyword
  • Show candidates → user confirms selection

3. Extract & Convert

Follow the Conversion section in the platform profile detected in Step 1. Save the result to 

{projectDir}/chats/.tmp/{timestamp}.yaml
.

4. Populate Metadata

The CLI auto-fills structural fields. The Skill's job is to fill in the human-facing metadata:

FieldCLI defaultAction
date
, 
sessionId
, 
model
, 
totalMessages
, 
totalTokens
, 
defaultShowProcess
Auto-filledReview only
visibility
private
Update to 
public
participants
Generic role names (
user
, 
assistant
)		Ask user for display names → rename keys; if no name is provided, use 
Human
/ 
Agent
for single-speaker or 
Human 1
/ 
Human 2
… for multi-speaker
title
'Session Export'
(generic)		Skim generated YAML → suggest → confirm
description
(absent)Write one-sentence summary → confirm
channel
(absent)Ask user — set to platform name (e.g. 
discord
) if applicable; omit otherwise
cover
(absent)Skip (user adds custom OG image URL manually later)
tags
(absent)Skip (user adds manually later)

5. Redact

Review the generated YAML and replace sensitive information with 

[REDACTED]
:

  • API keys, tokens, passwords → 
    [REDACTED]
  • File paths with usernames (
    /Users/xxx
    → 
    ~
    )
  • Email addresses, phone numbers → 
    [REDACTED]
  • Internal URLs and private IPs → 
    [REDACTED]

Do not silently remove content — always substitute 

[REDACTED]
so the reader can see that content was present but withheld.

6. Confirm & Save

  • Suggest filename: 
    {YYYYMMDD}-{topic}.yaml
  • Show preview → user confirms or modifies topic/filename
  • Before moving the file, create a dedicated branch (see below — required even if the user does not publish yet): 
    cd {projectDir}
git checkout main
git pull origin main
git checkout -b chat/{YYYYMMDD}-{topic}
  • Move: 
    {projectDir}/chats/.tmp/{timestamp}.yaml
    → 
    {projectDir}/chats/{YYYYMMDD}-{topic}.yaml
  • Stage and commit immediately so the file is isolated on its own branch: 
    git add chats/{YYYYMMDD}-{topic}.yaml
git commit -m "docs: add {topic}"

Why create a branch here? Saving on the default branch risks mixing unrelated changes into a future PR. Always commit each chat file on its own dedicated branch.

Optional: Publish

Push the branch created in step 6 and open a PR. See references/publish.md. Only proceed after explicit user request.

Edge Cases