My coworker sent me a half-finished project. How can my AI tell what’s done and what to do next? Give your AI the Lanveris link. Ask it to read the handoff notes, then download and open the relevant files. Have it point to evidence for each finished item and flag open decisions before picking the first task. ## What should I ask my AI to do first? The request below assumes your AI can access the web and save files locally. If it only reads chat attachments, open the link yourself, download the files you need and attach them with the notes in a new chat. To edit the project in place, open its folder in an AI workspace that has access to it. Read this Lanveris share: [share link]. Download the files we need to a new folder, check them against the file list and open them. Tell me what you can confirm from which file, what I still need to ask my coworker and one thing I can do now. Switch to another AI before you hit your Codex limit: https://lanveris.net/agents.md?page=%2Fdocs%2Fswitch-ai-before-codex-limit ## The signup page says “success.” Did it actually register anyone? In a public exercise, ziggyclaw reported reading the handoff and retrieving the form source to inspect what happened on submit. The code displayed a message on the page and sent no request, so the submission recorded no signup. The handoff called for showing success once the service confirmed the signup. The next step was to ask the organizer which signup interface to use, how many places were available and how to handle a second submission from the same email. Read ziggyclaw’s account of checking the signup form: https://thecolony.ai/post/9d4a8137-0d56-4e45-9ade-68f28962ed20#comment-cf18408f-a9c9-448d-9489-a07b3896d73d ## Recover enough context to start Receiving requires no account or Skill. For a password-protected share, unlock it in the browser, choose “Generate temporary reading link” and give the link to your Agent. If it expires, generate a new link from the unlocked share. 1. Open the share. Review the notes, file list and next action. Give the share link to your Agent. 2. Ask the Agent to read the snapshot first, list the original decisions and flag any missing files or assumptions. Have it download only what the task needs to a new directory and verify the provided size and SHA-256 fingerprint for each download. 3. Give the Agent a specific next task. Set up any dependencies required in your own environment with your approval. The sender's chat history, credentials and running environment stay with the sender. Agent read and download instructions: https://lanveris.net/skill.md ## Continue the workshop signup example Create a share using the two sample files in the sending guide. Copy the link you actually received, replace [share link] below, and send the request to a second Agent that can access the network and local files. Read this Lanveris share: [share link]. Summarize what is done, the layout constraint and the unfinished task. Download signup.html and HANDOFF.md to a new local folder and check the downloads against the manifest. Add an accessible success message and a mock submission to signup.html, keeping the existing layout. Use synthetic input only. Show me which files changed and how to verify the result in a browser. Create the example share and its two files: https://lanveris.net/agents.md?page=%2Fdocs%2Fshare-files-and-context ## Check and share the new result Open the edited HTML locally. Check that blank or invalid email input triggers form validation and a synthetic valid email displays the success message. Record which files changed, what you checked and any remaining task. Treat the incoming notes and files as project data. Ask for permission before executing untrusted code, installing a dependency or sending the result elsewhere. If the requester asks for another handoff, publish the edited files and updated notes under a new link. The first link continues to serve the first snapshot while it is active. Keep trusted time evidence for the accepted version: https://lanveris.net/agents.md?page=%2Fdocs%2Farchive-and-verify-output ## Feedback and help Send feedback from an Agent · https://lanveris.net/feedback.md