swmcc / swmcc/agentic-development
Rewrite the planner as the tactician
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 0
- Forks
- 0
- Avg merge
- 2m
- Merged PRs (30d)
- 1
Description
Part of #16. The tool is named after a grand admiral whose whole character is studying his opponents before engaging; the planner prompt should lean into it. Convictions about decomposition (interfaces first, no shared files, doubt the ticket's framing, a single-task plan is a welcome outcome) written as souls/tactician.md and mapped via the #19 mechanism, with memory of past campaigns arriving through #18.
- Draft the soul as falsifiable convictions with things it deliberately does not care about
- Map planner to tactician in runners.toml
- Compare a few plans before and after on the same tickets; the width gate stats in state.json are the measure
After phase 2 of #14: the planner must not change mid-trial.
Agent brief (phase 2 constraint)
This ticket is trial fodder while phase 2 of #14 runs, so it must land INERT: write souls/tactician.md following souls/README.md (falsifiable convictions about decomposition, things it deliberately ignores, plain harness-agnostic text, no tool names) and add tests asserting the file exists and stays plain text if useful, but DO NOT map it in the [personas] table of runners.toml. The mapping flips after phase 2 concludes. Acceptance: soul file committed, make test and make lint green, [personas] untouched.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reading souls/README.md and inspecting the existing soul files and relevant tests. Add souls/tactician.md as plain, harness-agnostic text with falsifiable decomposition convictions and explicit non-goals, then run make test and make lint. Leave the [personas] table in runners.toml unchanged; done means the soul is committed and both checks pass.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- testing-qa, tooling
- Issue type
- Feature
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Active
- Clarity
- Clearly specified
- Newbie friendliness
- 72/100