memfork 0.3.0

Git for agent memory, shared by every AI tool on your machine: fork, merge, rewind and hand off work between agents over MCP.
Documentation
description = "Write tests for existing code: find what is untested, cover the main paths and the edges, prove the tests can fail"

# Fill in each empty `accept` once with a command that exits 0 when that task
# is done, such as your test command. An empty one is never run.

[[task]]
id = "survey"
title = "List what the code does and which parts have no tests"
detail = "Store the list as a fact with the source files, so it is flagged stale if the code changes."

[[task]]
id = "main"
title = "Test the main behaviour"
depends_on = ["survey"]

[[task]]
id = "edges"
title = "Test the edge cases and the errors"
depends_on = ["survey"]

[[task]]
id = "verify"
title = "Run the new tests, and check each one fails when the code it tests is broken"
depends_on = ["main", "edges"]
accept = ""