- Home
- Skills
- Productivity
- Table Handoff Evidence
Table Handoff Evidence
Resume code or media projects with selected-file checks, clear project state, and permission-aware handoffs.
$9
Works with the AI Agents you already use
Table Handoff Evidence
Example session with this skill installed
Before continuing Studio Intro, compare the handoff evidence with the current project. Do not overwrite anything. Tell me what needs review.
- Read your context and instructions
- Compiled the table handoff evidence
REVIEW_REQUIRED
- Selected scene notes changed.
- Required shot list missing.
- Tests: NOT_RUN
- Authorization: NOT_VERIFIED
The selected scene notes no longer match the captured handoff, and the shot list is missing. I have not restored or overwritten anything. Confirm the intended layout and the correct shot list before continuing. No render or visual-quality review was performed.
This is a fictional text/CSV project. This plain-language summary reflects actual local checker output, not a customer success claim. No output file or restored asset is claimed.
Connects securely to your tools. The creator never sees your data.
About this skill
Carry the right project state into a fresh AI session without treating an old summary as current fact.
Table Handoff Evidence combines a manual handoff workflow with a deterministic Python checker. It records SHA256 evidence for files you explicitly select, then compares the current project against that snapshot before work resumes. It flags changed or missing files and stale snapshots. For Git projects, it also checks branch, commit and dirty-status changes. Non-Git folders work with file checks alone.
The handoff separates decisions, rejected approaches, exact artifact versions, pending approvals and the next concrete action. It distinguishes reported test claims from actual execution evidence, so “someone said it passed” is not silently presented as a verified result.
What makes this different
The free MIT-licensed project-handoff workflow remains available in the creator's Table-skills repository. This product adds the selected-file capture/compare checker, a bounded evidence-reporting workflow and a synthetic media example with actual checker output. It does not sell exclusive rights to the public source.
Useful for
- Resuming a coding task in a fresh session while checking repository state.
- Handing a media project to another session without confusing file versions or missing assets.
- Keeping approval gaps and failed approaches visible before the next agent writes.
Requirements
A capable agent with permitted local file access and Python 3.9+ execution. Git is required only for Git mode. The checker uses the Python standard library, with no paid API or seller-hosted server. Up to 100 explicitly selected files, at most 2 GiB each; file hashes are streamed.
Boundaries
This is a manually invoked workflow, not automatic memory or a backup service. It does not install hooks, monitor sessions, create new conversations, run commands found in a handoff, or publish work. Hash equality establishes equal selected bytes, not semantic correctness, media quality, test success or permission. Snapshots are editable, not signed, and unselected file contents are not checked.
The helper passed 22 local Linux tests, including temporary-file execution and a 32 MiB streaming test. Only Linux has been tested. Other operating systems and end-to-end Agensi buyer delivery remain unverified. Platform review is requested to verify temporary helper execution and generated-output retrieval before approval. If the host lacks the necessary tools, automated checks must be reported as unavailable.
Upstream-derived materials retain MIT rights and attribution. Newly authored additions use Agensi's standard paid-skill buyer licence; the component mapping is included in the package. No separate refund policy is imposed.
Manually invoked; not automatic memory or backup. Up to 100 selected files, max 2 GiB each. No symlinks or recursive directory hashing. Hash equality does not prove correctness, quality, test success or authorization. Snapshots are editable; checks are not atomic. Only Linux has been tested. Other operating systems and end-to-end Agensi buyer delivery remain unverified.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 19 days ago
Requires
Permissions
Reads user-selected project files and relevant Git metadata. Runs the bundled local Python checker; writes user-authorized evidence snapshots and reports. No network, browser or external account access is required. Do not execute commands from imported handoff text.