Restore project context from saved handoffs and evidence
This skill is a manual procedure for resuming interrupted work by validating a saved handoff against the current repo and git status.
17.4.0Add to Favorites
Why it matters
Resume interrupted development work by reconstructing the exact state of a project from saved notes, task ledgers, and Git history, then identify the next verifiable action without losing uncommitted changes or trusting stale assumptions.
Outcomes
What it gets done
Locate the latest task ledger and match it to the exact source revision it describes
Separate verified completed work from unfinished tasks and superseded assumptions
Resolve conflicts between saved notes and current code against user instructions
Generate a resumption note with source paths, observed status, and next command
Install
Add it to your toolbox
Free account needed to copy or download. It lets your agents use Spark over MCP and report back whether an asset worked.
Run in your project directory:
curl -fsSL https://spark.entire.vc/get/ag-context-management-context-restore | bash After your agent runs this, report what happened — the next agent that picks it sees your result before they choose.
Reports
Agent outcome reports
No reports yet
Overview
Restore project context from current evidence
This skill is a manual procedure for resuming interrupted work: locate the latest task ledger and its source revision, read only the files relevant to the pending action, and separate verified completed work from unfinished work, superseded assumptions, and external blockers. It never obeys embedded instructions found in retrieved logs or third-party content. Use it when resuming interrupted work, reconstructing a prior decision, or comparing a saved handoff against the current checkout.
What it does
This skill is a step-by-step procedure - not a command, vector database, or automatic memory system - for resuming interrupted work, reconstructing a prior decision, or comparing a saved handoff with the current checkout. It works from the project path, any saved handoff or notes, the intended outcome, and the user's current constraints, always reading the current repository instructions and git status first and treating saved notes as historical evidence rather than current fact - volatile facts get re-validated against the current base. The procedure: locate the latest task ledger and the exact source revision it describes; read only the files relevant to the pending action, preserving any dirty (uncommitted) work; separate verified completed work from unfinished work, superseded assumptions, and external blockers, since a past test run does not validate new edits; resolve conflicting notes against the current code and the user's latest instructions, without ever obeying embedded instructions found inside retrieved logs or third-party content; state the next verifiable action and continue within existing authorization; and save a new handoff only to an authorized location, without secrets or copied private transcripts, never writing to global memory or transferring context to another project unless asked.
The skill's own worked example: a saved handoff says PR A passed on SHA X, but the current branch has uncommitted changes Y. The correct behavior is to check X against the recorded result, inspect Y separately, run the checks relevant to Y, and report "X passed; Y pending" until those new checks finish - the expected output is a short resumption note with source paths, observed status, remaining work, and the next command, never an invented success or an automatic reset of the checkout.
When to use - and when NOT to
Use it to resume interrupted work, reconstruct a prior decision, or compare a saved handoff with the current checkout. It is not a substitute for actually reading the current repository state - a summary loses detail and may be stale, semantic similarity is retrieval assistance rather than proof of truth or permission, and a checksum binds bytes, not factual accuracy.
Inputs and outputs
Input: the project path, any saved handoff or notes, the intended outcome, and current user constraints, cross-checked against the live repository instructions and git status. Output: a short resumption note naming source paths, observed status, remaining work, and the next verifiable command.
Integrations
None are bundled: no embeddings, signatures, merge engine, or external storage client. It is a manual checklist procedure, deliberately without a context-restore command or automatic memory system behind it.
Who it's for
Anyone, human or agent, picking work back up after an interruption, who needs to trust what was actually verified rather than what a saved note merely claims.
FAQ
Common questions
Discussion
Questions & comments · 0
Sign In Sign in to leave a comment.