Uploads agent-generated artifacts (specs, plans, learnings) to the streamlit.wiki for sharing via PR comments. Use when you have agent artifacts to share with reviewers.
npx skills add https://github.com/streamlit/streamlit --skill sharing-pr-agent-artifacts
Uploads intermediate files (implementation plans, specs, learnings, explorations, architecture diagrams) to streamlit.wiki so they can be linked in PR descriptions and comments. Use this for agent-generated artifacts that are useful for reviewers but don't belong in the main repo.
Important: The wiki repo uses master as its default branch. Always push to master, never create other branches.
Follow the wiki guidelines: After cloning/updating the wiki repo, read and follow agent-wiki/AGENTS.md for the authoritative rules on what to upload and how to organize files.
Public URL pattern:
https://issues.streamlit.app/agent_wiki_explorer?file=<relative-path>
Example: https://issues.streamlit.app/agent_wiki_explorer?file=pull-requests/12345/implementation-plan.md
gh pr view --json number --jq '.number'
If no PR exists, stop and inform the user.
Check if the wiki is already cloned:
if [ -d "agent-wiki/.git" ]; then
# Use subshell to avoid directory navigation issues if pull fails
(cd agent-wiki && git checkout master && git pull origin master)
else
git clone https://github.com/streamlit/streamlit.wiki.git agent-wiki
fi
mkdir -p agent-wiki/pull-requests/<pr-number>
Common locations (check these first):
work-tmp/ — temporary working files, drafts, exploration notesspecs/ — untracked product/tech specs created for this PRWhat to include:
.md).png, .svg)Always exclude:
> Warning: The wiki is a public repo. Never upload confidential information, internal sensitive discussions, or anything that shouldn't be publicly visible. Since Streamlit is open-source, almost all tech and product related planning documents are fine to share publicly.
git ls-files to check).env, *credentials*, *secret*, *token*, *api_key*)Copy selected files:
cp <file> agent-wiki/pull-requests/<pr-number>/
# Use subshell to avoid directory navigation issues
(
cd agent-wiki
git checkout master
git add pull-requests/
git commit -m "Add artifacts for PR #<pr-number>"
git pull --rebase origin master
git push origin master
)
If push fails due to conflicts:
git pull --rebase origin mastergit rebase --continueNever use --force.
If new files were added, post a summary comment to the PR:
gh pr comment <pr-number> --body "$(cat <<'EOF'
### Added Agent Docs
- [implementation-plan.md](https://issues.streamlit.app/agent_wiki_explorer?file=pull-requests/<pr-number>/implementation-plan.md): Main technical implementation plan
- [exploration.md](https://issues.streamlit.app/agent_wiki_explorer?file=pull-requests/<pr-number>/exploration.md): Initial research and design exploration
EOF
)"
Include only top-level documents relevant to reviewers. Assets embedded in those documents (e.g., images) don't need separate entries.
For documents useful beyond a single PR (deep dives, architecture explorations, research findings), use the references/ directory instead of pull-requests/.
# Create file with date prefix
mkdir -p agent-wiki/references
cp <file> agent-wiki/references/YYYY-MM-DD-<description>.md
Example: 2024-03-15-widget-state-deep-dive.md
Use the same commit/push workflow as PR artifacts (step 5), substituting references/ for pull-requests/ in both the git add command and the commit message (e.g., "Add reference: YYYY-MM-DD-<description>").
Public URL: https://issues.streamlit.app/agent_wiki_explorer?file=references/YYYY-MM-DD-<description>.md
agent-wiki/ directory is gitignored and persists across sessionsmaster — never create feature branches or use --forceCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.
Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.
Replace with description of the skill and when Claude should use it.
Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies
This skill should be used when the user wants to "create a skill", "add a skill to plugin", "write a new skill", "improve skill description", "organize skill content", or needs guidance on skill structure, progressive disclosure, or skill development best practices for Claude Code plugins.
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.
Use when creating new skills, editing existing skills, or verifying skills work before deployment
Take streamlit/sharing-pr-agent-artifacts from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.