Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion ARCHITECTURE.md
Original file line number Diff line number Diff line change
Expand Up @@ -48,7 +48,7 @@ Groundwork ships skills from two maintenance locations, but inventory is unified

Skills maintained in this repository are listed in `skills/skills.toml` with local paths under `skills/`. These skills define the pipeline's structure — what stages exist, what handoff contracts connect them, and what cognitive discipline the pipeline enforces.

Skills maintained upstream (from [obra/superpowers](https://github.com/obra/superpowers)) are listed in the same manifest with pinned commits and fetched at install time. They fill the execution phase — debugging and subagent orchestration — where high-quality implementations already exist.
Skills maintained upstream (from [obra/superpowers](https://github.com/obra/superpowers)) are listed in the same manifest with pinned commits and fetched at install time. They fill the execution phase — subagent orchestration — where high-quality implementations already exist.

Curated skills are pinned to a specific commit. They are not forked, vendored, or modified. Integration happens through documentation: WORKFLOW.md defines handoff rules that connect curated skills to the pipeline's input/output contracts.

Expand Down
2 changes: 1 addition & 1 deletion README.md
Original file line number Diff line number Diff line change
Expand Up @@ -44,7 +44,7 @@ For the concise inventory and shipped order reference, see [`skills/skills.toml`
| `plan` | Decomposition | Unclear scope, design choices left to implementer |
| `test-first` | Execution | Implementation-first regressions |
| `subagent-driven-development` | Execution | Context drift in parallel work |
| `systematic-debugging` | Execution | Thrashing and symptom-fixing |
| `systematic-debugging` | Cross-cutting | Thrashing and symptom-fixing |
| `verification-before-completion` | Verification | False completion claims without evidence |
| `documentation` | Verification | Drifted docs, missing artifact updates |
| `propose` | Delivery | Manual ad-hoc commit/push/PR between implementation and merge |
Expand Down
14 changes: 13 additions & 1 deletion WORKFLOW.md
Original file line number Diff line number Diff line change
Expand Up @@ -36,7 +36,7 @@ Use `test-first` to implement each plan step through RED-GREEN-REFACTOR. Each RE

Use `subagent-driven-development` when the plan contains independent tasks that can run in parallel. It dispatches a fresh subagent per task with two-stage review (spec compliance, then code quality).

Use `systematic-debugging` when a test fails or behavior is unexpected. It finds root cause before proposing fixes — no symptom-patching.
Use `systematic-debugging` when a test fails or behavior is unexpected. It finds root cause before proposing fixes — no symptom-patching. Although listed here because most debugging occurs during execution, it is a cross-cutting discipline that fires at any pipeline stage when failures appear.

Code review is handled by CI/CD infrastructure, not a methodology skill. The pipeline requires review before landing but does not prescribe the mechanism.

Expand Down Expand Up @@ -89,6 +89,18 @@ Documentation is a cross-cutting communication discipline, not a pipeline phase.

For fail conditions and anti-divergence rules, see [docs/architecture/pipeline-contract.md](docs/architecture/pipeline-contract.md).

## Root-Cause Investigation Thread

Root-cause investigation is a cross-cutting discipline. It fires at any pipeline stage when a test fails, behavior is unexpected, or any failure occurs — before proposing fixes. See `systematic-debugging` for the full methodology.

### Handoff Rules

1. `systematic-debugging -> test-first`: once root cause is established, hand off to `test-first` fix-bug to write a failing test and implement the fix.
2. `systematic-debugging -> ground`: when the 3-fix escalation rule fires, invoke `ground` to re-examine architectural assumptions.
3. `systematic-debugging -> third-force`: when investigation reveals the root cause is operational friction (missing tool, broken config, stale convention), hand off to `third-force`.

For fail conditions and anti-divergence rules, see [docs/architecture/pipeline-contract.md](docs/architecture/pipeline-contract.md).

## Friction Resolution Thread

Friction resolution is a cross-cutting discipline. It fires at any pipeline stage when the operational environment impedes progress. See `third-force` for the full methodology.
Expand Down
2 changes: 1 addition & 1 deletion agents.toml
Original file line number Diff line number Diff line change
Expand Up @@ -19,5 +19,5 @@ using_groundwork = { gh = "pentaxis93/groundwork", path = "skills/using-groundwo
third_force = { gh = "pentaxis93/groundwork", path = "skills/third-force" }
subagent_driven_development = { gh = "obra/superpowers", path = "skills/subagent-driven-development", rev = "e4a2375cb705ca5800f0833528ce36a3faf9017a" }
test_first = { gh = "pentaxis93/groundwork", path = "skills/test-first" }
systematic_debugging = { gh = "obra/superpowers", path = "skills/systematic-debugging", rev = "e4a2375cb705ca5800f0833528ce36a3faf9017a" }
systematic_debugging = { gh = "pentaxis93/groundwork", path = "skills/systematic-debugging" }
verification_before_completion = { gh = "pentaxis93/groundwork", path = "skills/verification-before-completion" }
16 changes: 16 additions & 0 deletions docs/architecture/pipeline-contract.md
Original file line number Diff line number Diff line change
Expand Up @@ -87,6 +87,20 @@ Requirement:
Fail condition:
- `land` invoked on a branch with no PR, falling back to local merge and losing PR merge metadata.

### `systematic-debugging -> test-first`
Requirement:
- root cause is established before a fix is attempted through `test-first` fix-bug. Investigation output identifies the specific cause and the transition point where valid data becomes invalid.

Fail condition:
- fix-bug entered without root-cause analysis when the cause was unclear — agent proposed a fix from symptoms alone.

### `systematic-debugging -> ground`
Requirement:
- the 3-fix escalation rule triggers architectural re-examination via `ground` when 3 fix attempts have failed. The debugging scope ends and the architectural scope begins.

Fail condition:
- fourth fix attempt without questioning the architecture — agent continued to apply fixes past the point where the methodology indicated an architectural problem.

### `third-force -> documentation`
Requirement:
- structural fixes that change operational instructions are reflected in CLAUDE.md, CONTRIBUTING.md, or WORKFLOW.md.
Expand All @@ -111,6 +125,7 @@ Fail condition:
6. Do not land user-visible changes without a CHANGELOG entry.
7. Do not treat stale documentation as authoritative over code behavior.
8. Do not collapse the triad by routing around operational friction — resolve it structurally or file an issue.
9. Do not propose fixes without root-cause investigation when the cause is unclear.

## Quick Compliance Checklist

Expand All @@ -122,3 +137,4 @@ Fail condition:
- [ ] Documentation review completed before verification.
- [ ] User-facing changes include CHANGELOG entry.
- [ ] Documentation coverage status recorded at completion.
- [ ] Root-cause investigation completed before fix when cause was unclear.
3 changes: 1 addition & 2 deletions skills/skills.toml
Original file line number Diff line number Diff line change
Expand Up @@ -68,8 +68,7 @@ use_when = "Use when a verified plan contains independent tasks that can run in
name = "systematic-debugging"
path = "skills/systematic-debugging"
provider = "gh"
repo = "obra/superpowers"
rev = "e4a2375cb705ca5800f0833528ce36a3faf9017a"
repo = "pentaxis93/groundwork"
use_when = "Use when tests fail or behavior is unexpected and root cause is not yet established."

[[skills]]
Expand Down
37 changes: 37 additions & 0 deletions skills/systematic-debugging/LICENSE-UPSTREAM
Original file line number Diff line number Diff line change
@@ -0,0 +1,37 @@
This skill is adapted from the systematic-debugging skill in
obra/superpowers (https://github.com/obra/superpowers), pinned at commit
e4a2375. The original skill and its companion files (root-cause-tracing.md,
defense-in-depth.md, condition-based-waiting.md) are licensed under the MIT
License reproduced below.

Substantial portions of the original are preserved in this adaptation: the
Iron Law, the four-phase investigation model, the 3-fix architectural
escalation rule, the anti-rationalization table, and the red flags list.
These have been restructured as a cross-cutting discipline with vocabulary
normalization, language-agnostic examples, and bidirectional pipeline
integration with groundwork's test-first, verification, ground, and
third-force skills.

---

MIT License

Copyright (c) 2025 Jesse Vincent

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
Loading