Define immutable run manifests and typed artifacts
- Dominant language
- Python
- Stars
- 0
- Forks
- 0
- PR merge metrics
- No merged PRs in 30d
Description
## Parent and milestone
- Parent issue: #43
- Milestone: not scheduled; design may proceed after #42.
## Problem
Terminal logs and provider history do not establish which code, configuration,
runtime, plugins, environment, decisions, and validation produced a result.
## Outcome
Define an immutable run manifest and typed artifact model for reproducibility,
mobile review, incident investigation, and future web/mobile clients.
## Scope
- Record repository/base commit, workspace lease, effective configuration,
runtime/provider observations, plugin digests, sandbox profile, interactions,
validation, usage, final commits, and terminal outcome.
- Define typed diffs, commits, test reports, findings, checkpoints, logs, and PR
references with hashes and retention policy.
- Separate measured facts, estimates, agent claims, and user-authored metadata.
- Provide bounded view models and downloadable raw artifacts where authorized.
## Non-goals
- Storing hidden model reasoning.
- Treating generated summaries as authoritative evidence.
- Making Telegram formatting part of the artifact schema.
## Acceptance criteria
- [ ] A completed run can be attributed to exact inputs and verified outputs.
- [ ] Manifest replay cannot mutate recorded identity or evidence.
- [ ] Artifact hashes, provenance, retention, redaction, and access are defined.
- [ ] Telegram and future clients consume the same transport-neutral model.
- [ ] Partial, failed, cancelled, and recovered runs remain inspectable.
## Validation
- Fixture review for success, failure, cancellation, recovery, missing
artifact, redaction, and incompatible-schema cases.
## Dependencies
- #27 checkpoints, #41 control API, #44 execution environments, and #46 data
protection.
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by reading the dependent issues #27, #41, #44, and #46 to understand checkpoints, control API, execution environments, and data protection. Define the transport-neutral manifest and artifact model against the listed acceptance criteria, then review fixtures covering success, failure, cancellation, recovery, missing artifacts, redaction, and incompatible schemas.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- backend-api-design
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100