[Preset]: Add Autonomous Run Governance

[hindermath]

Preset ID

autonomous-run-governance

Preset Name

Autonomous Run Governance

Version

0.1.1

Description

Adds permission-bounded, evidence-first governance for autonomous Spec Kit delivery, convergence, resume, closeout, and retrospective learning.

Author

Thorsten Hindermann

Repository URL

https://github.com/hindermath/spec-kit-preset-autonomous-run-governance

Download URL

https://github.com/hindermath/spec-kit-preset-autonomous-run-governance/archive/refs/tags/v0.1.1.zip

Documentation URL

https://github.com/hindermath/spec-kit-preset-autonomous-run-governance/blob/main/README.md

License

MIT

Required Spec Kit Version

=0.8.3

Required Extensions (optional)

None

Templates Provided

• constitution-template - delivery authority, permission, and safety governance
• spec-template - applicability and scope boundaries
• plan-template - convergence, evidence, validation, resume, and closeout planning
• tasks-template - dependency-ordered autonomous delivery tasks
• autonomous-run-agent-guidance-addendum-template - portable agent guidance
• autonomous-runbook-template - permission-bounded execution runbook
• autonomous-run-evidence-template - convergence, validation, and delivery ledger
• autonomous-run-retrospective-template - classified workflow-learning evidence
• autonomous-run-readiness-checklist-template - local, PR, and merge-and-sync readiness
• autonomous-run-field-validation-summary - field-validation summary

Commands Provided

• speckit.autonomous - orchestrates a complete Spec Kit feature under explicit delivery authority
• speckit.autonomous-retrospective - extracts and classifies reusable learning from a completed run

Number of Scripts (optional)

0

Tags

autonomous, governance, evidence, permissions, retrospective

Key Features

• Uses LocalImplementation as the safe default and requires explicit current authorization for PublishPR or MergeAndSync.
• Drives Specify, repeated Clarify, checklists, Plan, Tasks, repeated Analyze, Implement, validation, delivery, and retrospective learning to evidence-based convergence.
• Keeps remote write, merge, bypass, cancellation, secret, and provider-administration authority explicit rather than inferred.
• Supports deterministic resume, non-recursive closeout, no-empty-PR rules, and independently reviewable evidence.
• Remains optional, stackable, provider-neutral, and independent of programming language, CI provider, branch versioning, and merge policy.
• Field-tested through six consecutive end-to-end TuiVision product features plus audit and closure runs. The public v0.1.1 ZIP was revalidated with Spec Kit 0.12.11; SHA-256: 517e3b254617d8545ab26cb816290a402fd8c94a52c603ce74916e1d51844032.
• Prior scope, safety, compatibility, and quality-adjusted throughput coordination: [Feature]: Add an optional Autonomous Run Governance preset for end-to-end Spec Kit delivery #3479

Testing Checklist

• Preset installs successfully via specify preset add
• Template resolution works correctly after installation
• Documentation is complete and accurate
• Tested on at least one real project

Submission Requirements

• Valid preset.yml manifest included
• Linked README (Documentation URL) explains how to use this preset and includes a valid specify preset add ... command (preferably specify preset add --from <download-url> using the exact download URL)
• LICENSE file included
• GitHub release created with version tag
• Preset ID follows naming conventions (lowercase-with-hyphens)

[hindermath]

Hi @mnriem, thank you for pointing me to the formal preset submission template in #3479. This issue was created through preset_submission.yml. The form currently applied enhancement, while the catalog workflow is triggered by preset-submission. Could you please add the preset-submission label when convenient so the automated validation can run?

[github-actions]

cc @github/spec-kit-maintainers — new catalog submission for review.

[github-actions]

Pull request created: #3501

Generated by Add Community Preset from Issue Submission · 254.5 AIC · ⌖ 15 AIC · ⊞ 33.8K

[hindermath]

Patch release and additional field validation

The published preset has now been updated to v0.1.2 based on the next TuiVision MergeAndSync field run.

• Release: https://github.com/hindermath/spec-kit-preset-autonomous-run-governance/releases/tag/v0.1.2
• Source PR: feat: release autonomous run governance v0.1.2 hindermath/spec-kit-preset-autonomous-run-governance#2
• Versioned ZIP: https://github.com/hindermath/spec-kit-preset-autonomous-run-governance/archive/refs/tags/v0.1.2.zip
• ZIP SHA-256: 6e4012895edf0fc9f8e4bd7795f02397190a28411034f9211cfaba5df899e2d0
• TuiVision adoption: chore: adopt autonomous run governance v0.1.2 hindermath/TuiVision#73

This patch adds two provider-neutral evidence safeguards:

1. validate the exact intended delivery candidate, including untracked-file coverage through git diff --cached --check, staged-path reconciliation, and a non-mutating temporary-index boundary for LocalImplementation;
2. map every acceptance gate to the workflow, job, runner or platform, and command that actually executed it, so a green aggregate or platform-named tooling job cannot stand in for missing technical proof.

Validation used Spec Kit 0.12.11 and the complete seven-preset stack. Codex, Claude, Copilot skills mode, and Antigravity each generated exactly one autonomous skill and one retrospective skill. Disable/enable/remove/reinstall, template resolution, public tag-ZIP installation, checksum verification, local-only index preservation, and authorized staged-candidate checks all passed. TuiVision then adopted v0.1.2 from the public ZIP; its full PR checks, DocFX/A11Y path, Claude review, and agent-parity checks passed.

The submission scope is unchanged: the preset remains optional, LocalImplementation remains the safe default, and installation grants no commit, push, merge, bypass, or provider-administration authority. The generated catalog PR #3501 can continue to use the existing submission while this comment supplies the newer release evidence.

[mnriem]

Thank you!

[hindermath]

Patch-release follow-up after catalog acceptance:

• autonomous-run-governance v0.1.4 is published: https://github.com/hindermath/spec-kit-preset-autonomous-run-governance/releases/tag/v0.1.4
• Exact tag ZIP: https://github.com/hindermath/spec-kit-preset-autonomous-run-governance/archive/refs/tags/v0.1.4.zip
• SHA-256: da667e2fd3fc5ccf0a29f7fd078d9f030f50ba267f659fc5b31bc000b59767e0
• Home Baseline PRs Add ability to set custom prefix to the / commands when running init #62/[Feature Request] Add support for using AugmentCode (AugmentCode extension or Auggie CLI tool) #63 and public preset PRs Update README.md #3/doco(spec-driven): Fix small typo in spec-driven.md #4 are merged.
• TuiVision adopted the exact public ZIP in chore: adopt autonomous governance v0.1.4 hindermath/TuiVision#76 (merge 9cc6f463916643d82c27392d32c96361e49366a8).

Field dogfooding found that Spec Kit ZIP installation copies the Bash validator without its Git executable mode (0644). v0.1.4 therefore standardizes bash <validator.sh> and pwsh -NoProfile -File <validator.ps1> while retaining the v0.1.3 machine-readable exact-head gate contract, priority 70, 14 contributions, and all permission boundaries.

Validation covered the seven-preset composition, unique agent surfaces, exact payload parity, equivalent Bash/PowerShell fixtures, DocFX/A11Y/text checks, and the exact TuiVision PR head. Validator success remains read-only and grants no remote or merge authority.