taskledger 0.6.5__tar.gz → 0.6.7__tar.gz

{taskledger-0.6.5 → taskledger-0.6.7}/.ledger/releaseledger/data/ledgers/main/events/events.jsonl

@@ -38,3 +38,15 @@
 {"data":{"updated_rows":3},"event":"audit.updated","event_id":"event-0038","record_revisions":{"commit_audit_sheet":2},"release_version":"v0.6.5","schema_version":2}
 {"data":{"entry_ids":["entry-0001","entry-0002","entry-0003"],"synced_audit_rows":3},"event":"entry.batch_added","event_id":"event-0039","record_revisions":{"commit_audit_sheet":3,"entry:v0.6.5/entry-0001":1,"entry:v0.6.5/entry-0002":1,"entry:v0.6.5/entry-0003":1,"release:v0.6.5":3},"release_version":"v0.6.5","schema_version":2}
 {"data":{"released_at":"2026-08-22"},"event":"release.finalized","event_id":"event-0040","record_revisions":{"release:v0.6.5":4},"release_version":"v0.6.5","schema_version":2}
+{"data":{"status":"planned"},"event":"release.created","event_id":"event-0041","record_revisions":{"release:v0.6.6":1},"release_version":"v0.6.6","schema_version":2}
+{"data":{"fields":["changelog_file","git_base_ref","git_base_sha","git_commit_count","git_head_ref","git_head_sha","git_range"]},"event":"release.updated","event_id":"event-0042","record_revisions":{"release:v0.6.6":2},"release_version":"v0.6.6","schema_version":2}
+{"data":{"row_count":1},"event":"audit.created","event_id":"event-0043","record_revisions":{"commit_audit_sheet":1},"release_version":"v0.6.6","schema_version":2}
+{"data":{"updated_rows":1},"event":"audit.updated","event_id":"event-0044","record_revisions":{"commit_audit_sheet":2},"release_version":"v0.6.6","schema_version":2}
+{"data":{"entry_ids":["entry-0001"],"synced_audit_rows":1},"event":"entry.batch_added","event_id":"event-0045","record_revisions":{"commit_audit_sheet":3,"entry:v0.6.6/entry-0001":1,"release:v0.6.6":3},"release_version":"v0.6.6","schema_version":2}
+{"data":{"released_at":"2026-09-04"},"event":"release.finalized","event_id":"event-0046","record_revisions":{"release:v0.6.6":4},"release_version":"v0.6.6","schema_version":2}
+{"data":{"status":"planned"},"event":"release.created","event_id":"event-0047","record_revisions":{"release:v0.6.7":1},"release_version":"v0.6.7","schema_version":2}
+{"data":{"fields":["git_base_ref","git_base_sha","git_commit_count","git_head_ref","git_head_sha","git_range"]},"event":"release.updated","event_id":"event-0048","record_revisions":{"release:v0.6.7":2},"release_version":"v0.6.7","schema_version":2}
+{"data":{"row_count":1},"event":"audit.created","event_id":"event-0049","record_revisions":{"commit_audit_sheet":1},"release_version":"v0.6.7","schema_version":2}
+{"data":{"updated_rows":1},"event":"audit.updated","event_id":"event-0050","record_revisions":{"commit_audit_sheet":2},"release_version":"v0.6.7","schema_version":2}
+{"data":{"entry_ids":["entry-0001"],"synced_audit_rows":1},"event":"entry.batch_added","event_id":"event-0051","record_revisions":{"commit_audit_sheet":3,"entry:v0.6.7/entry-0001":1,"release:v0.6.7":3},"release_version":"v0.6.7","schema_version":2}
+{"data":{"released_at":"2026-09-09"},"event":"release.finalized","event_id":"event-0052","record_revisions":{"release:v0.6.7":4},"release_version":"v0.6.7","schema_version":2}

taskledger-0.6.7/.ledger/releaseledger/data/ledgers/main/releases/v0.6.6/audit/commit-audit.yaml

@@ -0,0 +1,87 @@
+schema_version: 1
+object_type: commit_audit_sheet
+versioning:
+  schema_version: 1
+  revision: 3
+release_version: v0.6.6
+git_base_ref: v0.6.5
+git_base_sha: 412b465c3900d28b9f0cc8eb2ef78f10bb8340b3
+git_head_ref: HEAD
+git_head_sha: 307d68a1cbbcc1639bd18affdeefb69064af3a6e
+git_range: 412b465c3900d28b9f0cc8eb2ef78f10bb8340b3..307d68a1cbbcc1639bd18affdeefb69064af3a6e
+commit_count: 1
+rows:
+  - sha: 307d68a1cbbcc1639bd18affdeefb69064af3a6e
+    source_ref: git:307d68a1cbbcc1639bd18affdeefb69064af3a6e
+    short_sha: 307d68a
+    changed_paths:
+      - API.md
+      - README.md
+      - docs/api.md
+      - docs/command_contract.md
+      - docs/full_task_cycle.md
+      - docs/public_surface.md
+      - docs/service_boundary_whitelist.md
+      - docs/usage.md
+      - skills/taskledger/SKILL.md
+      - taskledger/api/task_runs.py
+      - taskledger/cli_validate.py
+      - taskledger/command_inventory.py
+      - taskledger/services/agent_transcripts.py
+      - taskledger/services/handoff.py
+      - taskledger/services/tasks.py
+      - taskledger/services/validation_flow.py
+      - taskledger/services/worker_context.py
+      - taskledger/services/workflow_guidance.py
+      - tests/test_agent_command_logging.py
+      - tests/test_command_inventory.py
+      - tests/test_docs_and_skill.py
+      - tests/test_service_boundaries.py
+      - tests/test_validation_command.py
+      - tests/test_worker_pipeline_context.py
+      - tests/test_workflow_guidance.py
+    stats:
+      files_changed: 0
+      insertions: 475
+      deletions: 11
+    inspected: true
+    inspected_paths:
+      - API.md
+      - README.md
+      - docs/api.md
+      - docs/command_contract.md
+      - docs/full_task_cycle.md
+      - docs/public_surface.md
+      - docs/service_boundary_whitelist.md
+      - docs/usage.md
+      - skills/taskledger/SKILL.md
+      - taskledger/api/task_runs.py
+      - taskledger/cli_validate.py
+      - taskledger/command_inventory.py
+      - taskledger/services/agent_transcripts.py
+      - taskledger/services/handoff.py
+      - taskledger/services/tasks.py
+      - taskledger/services/validation_flow.py
+      - taskledger/services/worker_context.py
+      - taskledger/services/workflow_guidance.py
+      - tests/test_agent_command_logging.py
+      - tests/test_command_inventory.py
+      - tests/test_docs_and_skill.py
+      - tests/test_service_boundaries.py
+      - tests/test_validation_command.py
+      - tests/test_worker_pipeline_context.py
+      - tests/test_workflow_guidance.py
+    observed_behavior:
+      Adds validate command to the CLI and public task-run API; during
+      an active validation run it executes managed argv, records cwd, exit status, stdout,
+      stderr, transcript, and large-output artifacts, while leaving criterion status
+      classification explicit. Adds validator guidance and documentation distinguishing
+      probe/setup failures from target-behavior failures, with focused regression coverage.
+    public_impact: public
+    decision: accepted
+    stale: false
+    evidence_subject: "feat: improved validation command features"
+    target_entry_id: entry-0001
+    notes:
+      Reviewed implementation, API, documentation, command inventory, and focused
+      tests.

taskledger-0.6.7/.ledger/releaseledger/data/ledgers/main/releases/v0.6.6/entries/entry-0001.md

@@ -0,0 +1,52 @@
+---
+schema_version: 2
+object_type: release_entry
+versioning:
+  schema_version: 1
+  revision: 1
+entry_id: entry-0001
+release_version: v0.6.6
+kind: added
+summary:
+  Added managed validation command evidence with durable output capture and
+  explicit probe-failure classification
+status: accepted
+audience: null
+scopes: []
+source_refs:
+  - git:307d68a1cbbcc1639bd18affdeefb69064af3a6e
+paths:
+  - API.md
+  - README.md
+  - docs/api.md
+  - docs/command_contract.md
+  - docs/full_task_cycle.md
+  - docs/public_surface.md
+  - docs/service_boundary_whitelist.md
+  - docs/usage.md
+  - skills/taskledger/SKILL.md
+  - taskledger/api/task_runs.py
+  - taskledger/cli_validate.py
+  - taskledger/command_inventory.py
+  - taskledger/services/agent_transcripts.py
+  - taskledger/services/handoff.py
+  - taskledger/services/tasks.py
+  - taskledger/services/validation_flow.py
+  - taskledger/services/worker_context.py
+  - taskledger/services/workflow_guidance.py
+  - tests/test_agent_command_logging.py
+  - tests/test_command_inventory.py
+  - tests/test_docs_and_skill.py
+  - tests/test_service_boundaries.py
+  - tests/test_validation_command.py
+  - tests/test_worker_pipeline_context.py
+  - tests/test_workflow_guidance.py
+issues: []
+prs: []
+sources:
+  - git:307d68a1cbbcc1639bd18affdeefb69064af3a6e
+contributors: []
+breaking: false
+internal: false
+order: 1
+---

taskledger-0.6.7/.ledger/releaseledger/data/ledgers/main/releases/v0.6.6/release.md

@@ -0,0 +1,27 @@
+---
+schema_version: 2
+object_type: release
+versioning:
+  schema_version: 1
+  revision: 4
+version: v0.6.6
+status: released
+history_state: curated
+title: null
+released_at: "2026-09-04"
+previous_version: v0.6.5
+cancel_reason: null
+superseded_by: null
+changelog_file: docs/changelog.md
+boundary_ref: null
+source_refs: []
+source_count: null
+entry_count: 1
+artifact_count: 0
+git_base_ref: v0.6.5
+git_base_sha: 412b465c3900d28b9f0cc8eb2ef78f10bb8340b3
+git_head_ref: HEAD
+git_head_sha: 307d68a1cbbcc1639bd18affdeefb69064af3a6e
+git_range: 412b465c3900d28b9f0cc8eb2ef78f10bb8340b3..307d68a1cbbcc1639bd18affdeefb69064af3a6e
+git_commit_count: 1
+---

taskledger-0.6.7/.ledger/releaseledger/data/ledgers/main/releases/v0.6.7/audit/commit-audit.yaml

@@ -0,0 +1,87 @@
+schema_version: 1
+object_type: commit_audit_sheet
+versioning:
+  schema_version: 1
+  revision: 3
+release_version: v0.6.7
+git_base_ref: v0.6.6
+git_base_sha: f7c915cf9d515ab1c15bdffeb73162340d7534df
+git_head_ref: HEAD
+git_head_sha: 562fc8796f821e312877f0c5b2aa09f5daddb138
+git_range: f7c915cf9d515ab1c15bdffeb73162340d7534df..562fc8796f821e312877f0c5b2aa09f5daddb138
+commit_count: 1
+rows:
+  - sha: 562fc8796f821e312877f0c5b2aa09f5daddb138
+    source_ref: git:562fc8796f821e312877f0c5b2aa09f5daddb138
+    short_sha: 562fc87
+    changed_paths:
+      - README.md
+      - docs/command_contract.md
+      - docs/sync.md
+      - docs/transfer.md
+      - docs/usage.md
+      - taskledger/api/config.py
+      - taskledger/exchange.py
+      - taskledger/services/agent_logging.py
+      - taskledger/services/change_tracking.py
+      - taskledger/services/doctor.py
+      - taskledger/services/doctor_checks/artifact_checks.py
+      - taskledger/services/git_sync.py
+      - taskledger/services/implementation_flow.py
+      - taskledger/services/tasks.py
+      - taskledger/services/validation_flow.py
+      - taskledger/storage/agent_logs.py
+      - taskledger/storage/artifact_policy.py
+      - taskledger/storage/project_config.py
+      - tests/test_agent_command_logging.py
+      - tests/test_artifact_policy.py
+      - tests/test_config_cli.py
+      - tests/test_doctor.py
+      - tests/test_project_root_config.py
+      - tests/test_sync_git.py
+      - tests/test_taskledger_v2_exchange.py
+      - tests/test_validation_command.py
+    stats:
+      files_changed: 0
+      insertions: 700
+      deletions: 33
+    inspected: true
+    inspected_paths:
+      - README.md
+      - docs/command_contract.md
+      - docs/sync.md
+      - docs/transfer.md
+      - docs/usage.md
+      - taskledger/api/config.py
+      - taskledger/exchange.py
+      - taskledger/services/agent_logging.py
+      - taskledger/services/change_tracking.py
+      - taskledger/services/doctor.py
+      - taskledger/services/doctor_checks/artifact_checks.py
+      - taskledger/services/git_sync.py
+      - taskledger/services/implementation_flow.py
+      - taskledger/services/tasks.py
+      - taskledger/services/validation_flow.py
+      - taskledger/storage/agent_logs.py
+      - taskledger/storage/artifact_policy.py
+      - taskledger/storage/project_config.py
+      - tests/test_agent_command_logging.py
+      - tests/test_artifact_policy.py
+      - tests/test_config_cli.py
+      - tests/test_doctor.py
+      - tests/test_project_root_config.py
+      - tests/test_sync_git.py
+      - tests/test_taskledger_v2_exchange.py
+      - tests/test_validation_command.py
+    observed_behavior:
+      Added a shared artifact-size policy that bounds stored command
+      output to a configurable limit capped at 20,000,000 bytes, preserves UTF-8-safe
+      head/tail excerpts with truncation metadata, rejects oversized existing artifacts
+      during doctor, export/import, and managed Git sync, and documents the policy with
+      regression coverage.
+    public_impact: public
+    decision: accepted
+    stale: false
+    evidence_subject: "feat: limit artifact file size"
+    target_entry_id: entry-0001
+    notes: ""

taskledger-0.6.7/.ledger/releaseledger/data/ledgers/main/releases/v0.6.7/entries/entry-0001.md

@@ -0,0 +1,53 @@
+---
+schema_version: 2
+object_type: release_entry
+versioning:
+  schema_version: 1
+  revision: 1
+entry_id: entry-0001
+release_version: v0.6.7
+kind: added
+summary:
+  Added configurable artifact limits with UTF-8-safe truncation across command
+  evidence, archives, doctor, and Git sync
+status: accepted
+audience: null
+scopes: []
+source_refs:
+  - git:562fc8796f821e312877f0c5b2aa09f5daddb138
+paths:
+  - README.md
+  - docs/command_contract.md
+  - docs/sync.md
+  - docs/transfer.md
+  - docs/usage.md
+  - taskledger/api/config.py
+  - taskledger/exchange.py
+  - taskledger/services/agent_logging.py
+  - taskledger/services/change_tracking.py
+  - taskledger/services/doctor.py
+  - taskledger/services/doctor_checks/artifact_checks.py
+  - taskledger/services/git_sync.py
+  - taskledger/services/implementation_flow.py
+  - taskledger/services/tasks.py
+  - taskledger/services/validation_flow.py
+  - taskledger/storage/agent_logs.py
+  - taskledger/storage/artifact_policy.py
+  - taskledger/storage/project_config.py
+  - tests/test_agent_command_logging.py
+  - tests/test_artifact_policy.py
+  - tests/test_config_cli.py
+  - tests/test_doctor.py
+  - tests/test_project_root_config.py
+  - tests/test_sync_git.py
+  - tests/test_taskledger_v2_exchange.py
+  - tests/test_validation_command.py
+issues: []
+prs: []
+sources:
+  - git:562fc8796f821e312877f0c5b2aa09f5daddb138
+contributors: []
+breaking: false
+internal: false
+order: 1
+---

taskledger-0.6.7/.ledger/releaseledger/data/ledgers/main/releases/v0.6.7/release.md

@@ -0,0 +1,27 @@
+---
+schema_version: 2
+object_type: release
+versioning:
+  schema_version: 1
+  revision: 4
+version: v0.6.7
+status: released
+history_state: curated
+title: Release v0.6.7
+released_at: "2026-09-09"
+previous_version: v0.6.6
+cancel_reason: null
+superseded_by: null
+changelog_file: null
+boundary_ref: null
+source_refs: []
+source_count: null
+entry_count: 1
+artifact_count: 0
+git_base_ref: v0.6.6
+git_base_sha: f7c915cf9d515ab1c15bdffeb73162340d7534df
+git_head_ref: HEAD
+git_head_sha: 562fc8796f821e312877f0c5b2aa09f5daddb138
+git_range: f7c915cf9d515ab1c15bdffeb73162340d7534df..562fc8796f821e312877f0c5b2aa09f5daddb138
+git_commit_count: 1
+---

{taskledger-0.6.5 → taskledger-0.6.7}/API.md

@@ -168,6 +168,7 @@ payload = plan_guidance(Path.cwd(), "task-0001")
 - `add_change`
 - `scan_changes`
 - `run_implementation_command`
+- `run_validation_command`
 - `finish_implementation`
 - `show_task_run`
 - `start_validation`

{taskledger-0.6.5/taskledger.egg-info → taskledger-0.6.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: taskledger
-Version: 0.6.5
+Version: 0.6.7
 Summary: Durable project-state storage and CLI for coding workflows
 Author: Taskledger Contributors
 Maintainer: Holger Nahrstaedt
@@ -81,7 +81,7 @@ plan start -> plan guidance -> plan template -> plan check -> plan upsert -> pla
 question add | question add-many | question answer | question answer-many | question status | question answers
 todo next | todo show | todo done | todo status
 implement start | implement resume | implement change | implement scan-changes | implement finish
-validate start | validate status | validate check | validate finish
+validate start | validate status | validate command | validate check | validate finish
 review record
 handoff create | handoff show | handoff claim | handoff close
 ```
@@ -133,6 +133,8 @@ taskledger file link specs/behavior/features/checkout/payment.feature --kind doc
 taskledger validate check --criterion ac-0001 --status pass --evidence "pytest tests/test_checkout_payment.py::test_payment_flow"
 ```
 
+Validation command failures must be classified before recording a criterion failure. A non-zero command is not automatically a failed acceptance criterion: distinguish a genuine target-behavior failure from a malformed probe or setup failure, repair the probe, and rerun it. Prefer `taskledger validate command -- ...` for durable command evidence, and normalize compared paths against the same explicit project root.
+
 ### Which read command to use
 
 | Need                       | Command                                             |
@@ -835,6 +837,12 @@ Import safety still relies on `project_uuid`, not the name/slug.
 - `--no-include-bodies` strips record body text (`body` / `context_body`) from exported payloads.
 - `--include-run-artifacts` embeds task and agent-log artifact files under `artifacts/` in the archive.
 
+## Artifact file-size policy
+
+Taskledger-owned artifact files are bounded to prevent command output from creating very large ledger blobs. The default and hard ceiling is 20,000,000 bytes per artifact; configure a lower `artifact_max_bytes` value when needed. Oversized command evidence keeps a UTF-8-safe head and tail with an explicit truncation marker, while command execution and exit-code semantics remain unchanged.
+
+Archive import/export applies the same policy, and `taskledger doctor` reports legacy oversized artifacts without reading their contents. `taskledger sync git push` performs a preflight before committing or pushing. Taskledger does not use Git LFS or rewrite Git history; repair existing blobs separately.
+
 Cross-machine imports preserve durable task/run data, but imported runtime locks
 are quarantined by default. After importing an in-progress implementation,
 run:

{taskledger-0.6.5 → taskledger-0.6.7}/README.md

@@ -28,7 +28,7 @@ plan start -> plan guidance -> plan template -> plan check -> plan upsert -> pla
 question add | question add-many | question answer | question answer-many | question status | question answers
 todo next | todo show | todo done | todo status
 implement start | implement resume | implement change | implement scan-changes | implement finish
-validate start | validate status | validate check | validate finish
+validate start | validate status | validate command | validate check | validate finish
 review record
 handoff create | handoff show | handoff claim | handoff close
 ```
@@ -80,6 +80,8 @@ taskledger file link specs/behavior/features/checkout/payment.feature --kind doc
 taskledger validate check --criterion ac-0001 --status pass --evidence "pytest tests/test_checkout_payment.py::test_payment_flow"
 ```
 
+Validation command failures must be classified before recording a criterion failure. A non-zero command is not automatically a failed acceptance criterion: distinguish a genuine target-behavior failure from a malformed probe or setup failure, repair the probe, and rerun it. Prefer `taskledger validate command -- ...` for durable command evidence, and normalize compared paths against the same explicit project root.
+
 ### Which read command to use
 
 | Need                       | Command                                             |
@@ -782,6 +784,12 @@ Import safety still relies on `project_uuid`, not the name/slug.
 - `--no-include-bodies` strips record body text (`body` / `context_body`) from exported payloads.
 - `--include-run-artifacts` embeds task and agent-log artifact files under `artifacts/` in the archive.
 
+## Artifact file-size policy
+
+Taskledger-owned artifact files are bounded to prevent command output from creating very large ledger blobs. The default and hard ceiling is 20,000,000 bytes per artifact; configure a lower `artifact_max_bytes` value when needed. Oversized command evidence keeps a UTF-8-safe head and tail with an explicit truncation marker, while command execution and exit-code semantics remain unchanged.
+
+Archive import/export applies the same policy, and `taskledger doctor` reports legacy oversized artifacts without reading their contents. `taskledger sync git push` performs a preflight before committing or pushing. Taskledger does not use Git LFS or rewrite Git history; repair existing blobs separately.
+
 Cross-machine imports preserve durable task/run data, but imported runtime locks
 are quarantined by default. After importing an in-progress implementation,
 run:

{taskledger-0.6.5 → taskledger-0.6.7}/docs/api.md

@@ -137,6 +137,7 @@ payload = plan_guidance(Path.cwd(), "task-0001")
 - `add_change`
 - `scan_changes`
 - `run_implementation_command`
+- `run_validation_command`
 - `finish_implementation`
 - `show_task_run`
 - `start_validation`

{taskledger-0.6.5 → taskledger-0.6.7}/docs/changelog.md

@@ -2,6 +2,22 @@
 
 > **Generated by releaseledger.** Manual edits may be overwritten. Update releaseledger entries and rebuild this file instead.
 
+## [v0.6.7] - 2026-09-09
+
+### Added
+
+- Added configurable artifact limits with UTF-8-safe truncation across command evidence, archives, doctor, and Git sync
+
+<!-- generated by releaseledger -->
+
+## [v0.6.6] - 2026-09-04
+
+### Added
+
+- Added managed validation command evidence with durable output capture and explicit probe-failure classification
+
+<!-- generated by releaseledger -->
+
 ## [v0.6.5] - 2026-08-22
 
 ### Added

{taskledger-0.6.5 → taskledger-0.6.7}/docs/command_contract.md

@@ -176,6 +176,9 @@ Rules:
 - Links are opaque. Taskledger does not parse or interpret linked files.
 - Validation evidence is recorded through generic `validate check`.
 
+- `validate command -- ...` records managed validation command evidence during an active validation run.
+- A non-zero validation command does not create a `ValidationCheck`; explicitly classify and record criteria with `validate check`.
+
 ## Archive import lock policy
 
 Archive imports support explicit lock handling:
@@ -448,6 +451,9 @@ shell. They execute the child from the CLI invocation directory, or the explicit
 `--root` directory, even when Taskledger discovers an ancestor workspace root for
 ledger state, artifacts, and logs.
 
+Managed command artifacts are bounded by the project `artifact_max_bytes` setting, defaulting to the 20,000,000-byte hard ceiling. Oversized persisted evidence is marked and retains both its beginning and end; only the durable artifact is truncated, not the child command's exit code or CLI stdout/stderr contract.
+`validate command` uses the same managed execution behavior during an active validation run, but records command evidence only. It never automatically maps a non-zero child exit to a criterion result.
+
 Managed commands preserve the caller environment, but resolve child executable lookup
 deterministically. A usable inherited `VIRTUAL_ENV` takes precedence. Otherwise
 Taskledger checks `<command-cwd>/.venv`, `<command-cwd>/venv`, then the workspace
@@ -474,6 +480,8 @@ taskledger plan command -- pytest tests/ -q
 taskledger plan command --allow-failure -- pytest tests/ -q
 taskledger implement command -- ruff check --config=.ruff.toml .
 taskledger implement command --allow-failure -- python -c "raise SystemExit(7)"
+taskledger validate command -- pytest tests/ -q
+taskledger validate command --allow-failure -- python -c "raise SystemExit(7)"
 ```
 
 ## Transcript review modes

{taskledger-0.6.5 → taskledger-0.6.7}/docs/full_task_cycle.md

@@ -264,6 +264,8 @@ taskledger validate show
 taskledger validate finish --result passed --summary "Parser fix validated with regression tests."
 ```
 
+Use `taskledger validate command -- ...` to preserve validation command cwd, output, exit status, and transcript evidence. A non-zero command is not automatically a failed acceptance criterion. If the command is a malformed probe or setup failure, correct it and rerun before recording criterion status. Only record `fail` after the target behavior was evaluated and failed.
+
 ### If validation finds a bug
 
 ```bash

{taskledger-0.6.5 → taskledger-0.6.7}/docs/public_surface.md

@@ -26,7 +26,7 @@ the normal planning/approval path for agents.
 - `question add`, `question add-many`, `question answer`, `question answer-many`, `question status`, `question answers`
 - `todo next`, `todo show`, `todo done`, `todo status`
 - `implement start`, `implement resume`, `implement change`, `implement scan-changes`, `implement finish`
-- `validate start`, `validate status`, `validate check`, `validate finish`
+- `validate start`, `validate status`, `validate command`, `validate check`, `validate finish`
 - `review record`
 - `handoff create`, `handoff show`, `handoff claim`, `handoff close`
 

{taskledger-0.6.5 → taskledger-0.6.7}/docs/service_boundary_whitelist.md

@@ -79,6 +79,8 @@ Current sanctioned imports:
   resolves actor/harness context for event metadata.
 - `taskledger/cli_implement.py:taskledger.services.agent_logging` —
   Implement command wrapper records managed-shell command failures.
+- `taskledger/cli_validate.py:taskledger.services.agent_logging` — Validation
+  command wrapper records managed-shell command failures.
 - `taskledger/cli_misc.py:taskledger.services.doctor` — Doctor commands
   still consume doctor service inspectors directly.
 - `taskledger/cli_pipeline.py:taskledger.services.handoff` — Pipeline

{taskledger-0.6.5 → taskledger-0.6.7}/docs/sync.md

@@ -81,6 +81,10 @@ report project-local vs outside-project dirty paths. `sync git push` commits
 repository-wide changes by design to match the standard Git workflow. Use
 `sync git cd` for advanced manual inspection or conflict resolution.
 
+## Artifact preflight
+
+`taskledger sync git push` checks Taskledger-owned artifact files before staging or committing. It refuses any existing file over the configured `artifact_max_bytes` limit and reports the paths, sizes, limit, and `taskledger doctor` remediation. Plain `git push` remains outside Taskledger's control, and Taskledger does not rewrite history or use Git LFS.
+
 For manual conflict resolution or debugging:
 
 ```bash

{taskledger-0.6.5 → taskledger-0.6.7}/docs/transfer.md

@@ -13,6 +13,12 @@ machines and harnesses.
   \- `project.ledger_ref` (exported ledger)
 - Optional run artifacts under `artifacts/` when `--include-run-artifacts` is set.
 
+## Artifact file-size policy
+
+Taskledger-owned artifact files default to a 20,000,000-byte hard ceiling. Projects may lower this with `artifact_max_bytes`, but cannot configure an unlimited or larger value. Oversized command output is stored as a UTF-8-safe head/tail excerpt with an explicit byte-count marker; the managed command's output and exit code are unchanged.
+
+Import/export enforce the same limit. Run `taskledger doctor` to find existing oversized files; `taskledger sync git push` refuses them before creating a commit. This does not rewrite existing Git history or use Git LFS.
+
 ## Filename policy
 
 When no output path is passed to `taskledger export`, taskledger writes into

{taskledger-0.6.5 → taskledger-0.6.7}/docs/usage.md

@@ -592,6 +592,8 @@ taskledger implement command -- ruff check --config=.ruff.toml .
 taskledger implement command --allow-failure -- python -c "raise SystemExit(7)"
 ```
 
+Use `taskledger validate command -- ...` for durable validation execution evidence. A non-zero command is not automatically a failed acceptance criterion. Classify whether the target behavior was evaluated before recording `fail`; repair malformed cwd, import, fixture, quoting, or path-normalization probes and rerun them first.
+
 When `[agent_logging].enabled = true` in `.ledger/taskledger/config.toml`, `taskledger`
 records CLI invocations and managed command outputs. Both managed wrappers execute
 the exact argv after `--` without a shell and from the CLI invocation directory, or
@@ -599,8 +601,8 @@ the explicit `--root` directory. Taskledger may discover an ancestor workspace f
 ledger state, but that does not change the child's cwd. Human mode displays captured
 stdout/stderr before the summary. JSON mode contains stdout/stderr and cwd only in
 the result object. Keep logging opt-in because stdout/stderr may contain sensitive
-data. Route task-relevant commands through `plan command` and `implement command`
-so their output is included in task transcripts and reports.
+data. Route task-relevant commands through `plan command`, `implement command`, and
+`validate command` so their output is included in task transcripts and reports.
 
 Managed commands use a usable inherited Python virtualenv first, then a valid
 `.venv` or `venv` in the command directory or workspace, preferring the command
@@ -760,6 +762,9 @@ Export include flags are content-affecting:
 - `--no-include-bodies` removes record body fields from the exported payload.
 - `--include-run-artifacts` embeds artifact files under `artifacts/` in the archive.
 
+## Artifact file-size policy
+
+Taskledger-owned artifact files are limited to 20,000,000 bytes by default and may only be lowered with `artifact_max_bytes`. Oversized command evidence retains its head and tail and includes a deterministic truncation marker; it does not change command execution or exit-code behavior. Import/export, doctor, and Taskledger-managed Git push enforce the same policy. Existing Git history is not rewritten and Git LFS is not used.
 Cross-machine imports preserve durable task/run records but quarantine imported
 runtime locks by default. For an imported in-progress implementation, run:
 

{taskledger-0.6.5 → taskledger-0.6.7}/skills/taskledger/SKILL.md

@@ -40,7 +40,7 @@ plan start -> plan guidance -> plan template -> plan check -> plan upsert -> pla
 question add | question add-many | question answer | question answer-many | question status | question answers
 todo next | todo show | todo done | todo status
 implement start | implement resume | implement change | implement scan-changes | implement finish
-validate start | validate status | validate check | validate finish
+validate start | validate status | validate command | validate check | validate finish
 review record
 handoff create | handoff review | handoff show | handoff claim | handoff release | handoff retarget | handoff close
 ```
@@ -356,13 +356,22 @@ linked-file drift checks.
 1. `taskledger context --for validation --format markdown`
 2. `taskledger validate start`
 3. Check `taskledger validate status` to see current validation state and blockers.
-4. Run verification checks outside taskledger.
+4. Run verification with `taskledger validate command -- ...` whenever practical.
 5. Record criterion results: `taskledger validate check --criterion ac-0001 --status pass|fail|warn|not_run --evidence "..."`
 6. Optionally waive criteria with user authority: `taskledger validate waive --criterion ac-0001 --reason "..."`.
 7. Check `taskledger validate status` again to confirm all mandatory gates pass.
 8. Finish with `taskledger validate finish --result passed|failed|blocked --summary "..."`.
    - For passed runs, prefer a concise done-work summary that can be reused by release-note tooling.
 
+### Validation command failure classification
+
+A non-zero validation command is not automatically a failed acceptance criterion. Before `validate check --status fail`, determine whether the command actually evaluated the target behavior.
+
+- **Target failure:** the check executed correctly and the accepted behavior failed. Record the criterion as `fail`.
+- **probe/setup failure:** the validator command is invalid or could not reach the target behavior, for example because of cwd, imports, quoting, fixtures, or relative/absolute path representation. Correct the probe and rerun. Do not mark the criterion failed solely because the probe was malformed.
+- If durable state is needed while repairing a probe, record the criterion as `not_run` with `--details` explaining the probe problem.
+- For path-set comparisons, normalize both sides against the same explicit project root before equality checks.
+
 ### Recovery Rules
 
 - If validation fails, record the failure and do not hide it.
@@ -396,7 +405,7 @@ taskledger validate check --criterion ac-0001 --status pass --evidence "pytest t
 - Log every meaningful implementation change with `taskledger implement change`.
 - Record deviations from the approved plan with `taskledger implement deviation`.
 - Mark todos done with evidence: `taskledger todo done <todo-id> --evidence "pytest -q"`.
-- When transcript logging is enabled in project config, run task-relevant commands through `taskledger plan command -- ...` or `taskledger implement command -- ...` so stdout/stderr becomes durable transcript evidence. The managed result and shell log include the actual child cwd.
+- When transcript logging is enabled in project config, run task-relevant commands through `taskledger plan command -- ...`, `taskledger implement command -- ...`, or `taskledger validate command -- ...` so stdout/stderr becomes durable transcript evidence. The managed result and shell log include the actual child cwd.
 - Do not paste secrets into command arguments, evidence strings, or expected command output when transcript logging is enabled.
 - Use `taskledger handoff create --mode implementation|validation --intended-actor agent --intended-harness codex` when switching actor or harness.
 - Use `taskledger handoff claim handoff-0001` before continuing work from a handoff.

{taskledger-0.6.5 → taskledger-0.6.7}/taskledger/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '0.6.5'
-__version_tuple__ = version_tuple = (0, 6, 5)
+__version__ = version = '0.6.7'
+__version_tuple__ = version_tuple = (0, 6, 7)
 
-__commit_id__ = commit_id = 'g412b465c3'
+__commit_id__ = commit_id = 'gfa5b2b772'