Repository navigation
Refresh PStack workflows without model changes - #26
Merged
Merged
Conversation
Contributor
There was a problem hiding this comment.
🔵 Needs a closer look
It reworks a large set of agent-governing workflow documents with intentional behavioral shifts (e.g., restoring worktree-default in opening-a-pr.md against AGENTS.MD, removing upstream MIT license/attribution copies) that warrant human judgment beyond the one convention fix flagged here.
3 open findings
What changed in this PR
This PR refreshes the repo's bundled agent "skills" (PStack workflows) to upstream cursor/plugins pstack 0.15.9, keeping all local model/reasoning routing untouched. It is overwhelmingly a documentation/prose refresh of SKILL.md and playbook files, plus two small, verified helper-script tweaks. It fits the repo's established "mine upstream, adapt for Codex, keep Bram-local policy" flow tracked in docs/upstream-mined.md.
Changes:
- Adds three skills (
correct,benchmark-checklist,principle-explain-the-number) and wires the latter two intopoteto-modetriggers/principles; adds architecture red-flags and schema-first TypeScript examples. - Rewrites autopilot/babysit/opening-a-pr/perf/hillclimb playbooks (fresh-agent defaults, code-ready verification rounds, hourly quiet audit heartbeats, performance-mantra ordering) and trims prose across many skills.
- Removes per-skill
LICENSEcopies and_Source_banners; updatescheck-plan.mjsprogram markers andlog.shheader-write logic; records the port inCHANGELOG.mdanddocs/upstream-mined.md.
| File | Description |
|---|---|
| skills/*/LICENSE (23 files: unslop, technical-writing, tdd, swarm, show-me-your-work, reflect, principle-*, poteto-mode, interrogate, how, figure-it-out, blast-radius, architect, …) | Removes per-skill MIT LICENSE copies; provenance retained in docs |
| skills/correct/SKILL.md, skills/correct/agents/openai.yaml | New $correct skill; policy file missing interface: block |
| skills/benchmark-checklist/SKILL.md, agents/openai.yaml | New perf-vetting skill; policy file missing interface: block |
| skills/principle-explain-the-number/SKILL.md, agents/openai.yaml | New principle; policy file missing interface: block |
| skills/poteto-mode/SKILL.md | Adds benchmark/explain-number triggers, fresh-subagent defaults, todolist wording |
| skills/poteto-mode/scripts/check-plan.mjs | Simplifies PROGRAM_MARKERS to string-only (hourly audit heartbeat/status message) |
| skills/poteto-mode/playbooks/{autopilot-full,autopilot-stack,babysit,opening-a-pr,multi-phase-plan,shipping,perf-issue,hillclimb,feature,bug-fix,refactoring,pause-safely}.md | Workflow refresh: verification rounds, heartbeats, worktree/PR-tool wording, perf mantras |
| skills/poteto-mode/references/poteto-agent.md | Fresh-agent routing wording |
| skills/{reflect/references/*,interrogate/references/*,how/references/*,architect/*,blast-radius,show-me-your-work,figure-it-out,swarm,tdd,unslop,technical-writing,typescript-best-practices,principle-*}.md | Prose trims and Codex tool-name adaptations |
| skills/show-me-your-work/scripts/log.sh | Uses -s/>> so a stray header costs one line, not the rows |
| skills/architect/references/design-red-flags.md | Adds split-ownership/importable-internals/hand-synced-list red flags |
| CHANGELOG.md, docs/upstream-mined.md | Record the 0.15.9 port and local retentions |
🧠 Review effort: Balanced
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
Comment on lines
+1
to
+2
| policy: | ||
| allow_implicit_invocation: false |
Comment on lines
+1
to
+2
| policy: | ||
| allow_implicit_invocation: false |
Comment on lines
+1
to
+2
| policy: | ||
| allow_implicit_invocation: false |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Refresh PStack workflows through upstream 0.15.9 (
e43c7ee26e), keeping upstream wording except required Codex adaptations and preserving all existing models and reasoning settings. Adds Correct, Benchmark checklist, and Explain the Number; updates architecture screening, verification receipts, autopilot audits, plan/log helpers, and prompt wording. Retains the existing plan checker fence and mandatory-task guards, with red/green regression proof. Removes the requested local per-skill license copies and source banners; root LICENSE is unchanged.Validation: five focused plan-checker cases, 79 Bun tests, TypeScript typecheck, 109 skill YAML checks, three invocation-policy checks, skill-mirror tests, shell/Node syntax, docs inventory, and whitespace checks pass. The unchanged skill wording also has six prior native Codex fixture dry runs and 184 compatibility checks. Native todo-list UI, real PR/merge operations, and actual scheduler execution were not covered by those fixture runs.