No implementation currently clears the Snapshot + Claim threshold; this is not a claim that the product lacks the capability.
Inspect coverage boundary →MECHANISM · SAFETY · CORE
Workspace Path Boundary
Reject absolute paths and traversal before a file tool touches the filesystem.
1 lessons0 agent snapshots0 experiments
safetypath-boundaryproblem → policy → evidence
L0Intuitionpartial
L1Buildcomplete
L2Engineeringpartial
L3Architecturemissing
L4Researchmissing
DEPENDENCY GRAPH · READER PATH
What feeds it, and what does it unlock?
Arrows show learning and design dependencies, not runtime data flow. Follow any node to continue through the Atlas.
Prerequisites
path-boundaryWorkspace Path BoundaryCurrent research boundary
Reject absolute paths and traversal before a file tool touches the filesystem.
The structured record is in the Atlas, but the L0–L4 article has not passed content review.
AGENT MAPPING · EVIDENCE ONLY
Snapshot implementations
An implementation enters the map only when both a snapshot and claims exist; unknowns remain visible.
No implementation currently clears the Snapshot + Claim threshold; this is not a claim that the product lacks the capability.
Inspect coverage boundary →No implementation currently clears the Snapshot + Claim threshold; this is not a claim that the product lacks the capability.
Inspect coverage boundary →No implementation currently clears the Snapshot + Claim threshold; this is not a claim that the product lacks the capability.
Inspect coverage boundary →No implementation currently clears the Snapshot + Claim threshold; this is not a claim that the product lacks the capability.
Inspect coverage boundary →No implementation currently clears the Snapshot + Claim threshold; this is not a claim that the product lacks the capability.
Inspect coverage boundary →No implementation currently clears the Snapshot + Claim threshold; this is not a claim that the product lacks the capability.
Inspect coverage boundary →CLAIMS · EVIDENCE LEDGER0 RECORDS
No reviewed Agent claims
The teaching-harness reference does not automatically become an architectural claim about any vendor agent.
EXPERIMENTS · EXERCISES
What has actually been tested?
Formal experiments are separate from course exercises. Exercises can validate the reference implementation but cannot replace Native evidence from a real agent.
Formal experiment · not registered
No experiment record means untested, not passed. The course checks below cover only the deterministic reference harness.
s04-workspace-toolsobserve
Verify the workspace readThe Golden Trace contains only a fixture-relative path and a clean result.
python3 -m curriculum.golden verify s04-workspace-toolss04-workspace-toolsmodify
Challenge the path boundaryA ../ escape is rejected before any filesystem operation occurs.
python3 -m unittest curriculum.tests.test_vertical_slice.VerticalSliceTests.test_workspace_rejects_escape -v01
How should adapters represent symlink and mount visibility across operating systems?