gsd-codex-cli 1.20.0

package/.codex/prompts/gsd-plan-milestone-gaps.md

@@ -0,0 +1,43 @@
+---
+description: Create phases to close all gaps identified by milestone audit
+argument-hint: [none]
+---
+
+
+
+## Objective
+Create phases to close all gaps identified by milestone audit
+
+## Compatibility
+- Use .codex/skills/get-shit-done-codex semantics.
+- Treat upstream workflow as the source of truth.
+- Replace user-specific paths with workspace-relative paths (.claude/..., .planning/...).
+- Run engine commands through PowerShell: 
+node .claude/get-shit-done/bin/gsd-tools.js ...
+- Parse JSON with ConvertFrom-Json; parse key/value output when workflow uses KEY=value raw mode.
+- No jq / bash-only constructs.
+
+## Subagent lifecycle (required)
+
+- Translate each upstream `Task(...)` into `spawn_agent` -> `wait` -> `close_agent`.
+- Spawn only when the upstream workflow defines an agent role.
+- Use `.claude/agents/gsd-*.md` as role context for each spawned agent.
+- Do not advance workflow steps until wait and close complete.
+## Execution
+1. Parse [none] from the user input.
+2. Run init:
+node .claude/get-shit-done/bin/gsd-tools.js init milestone-op --raw
+
+3. Load .claude/get-shit-done/workflows/plan-milestone-gaps.md and execute it step-by-step.
+4. Translate each Task(...) in workflow into:
+   - spawn_agent with the matching role file context from .claude/agents/.
+   - wait for each spawned agent and apply returned output before moving forward.
+5. Preserve all gates and routing from upstream workflow.
+6. Preserve commit behavior using 
+node .claude/get-shit-done/bin/gsd-tools.js commit "message" --files ....
+7. If commit preflight fails (no git / no commit flag), proceed in read-only mode and report clearly.
+
+## Completion output
+- Summarize key artifacts created/updated and next recommended command.
+
+

package/.codex/prompts/gsd-plan-phase.md

@@ -0,0 +1,43 @@
+---
+description: Create detailed execution plan for a phase (PLAN.md) with verification loop
+argument-hint: "[phase] [--research] [--skip-research] [--gaps] [--skip-verify]"
+---
+
+
+
+## Objective
+Create detailed execution plan for a phase (PLAN.md) with verification loop
+
+## Compatibility
+- Use .codex/skills/get-shit-done-codex semantics.
+- Treat upstream workflow as the source of truth.
+- Replace user-specific paths with workspace-relative paths (.claude/..., .planning/...).
+- Run engine commands through PowerShell: 
+node .claude/get-shit-done/bin/gsd-tools.js ...
+- Parse JSON with ConvertFrom-Json; parse key/value output when workflow uses KEY=value raw mode.
+- No jq / bash-only constructs.
+
+## Subagent lifecycle (required)
+
+- Translate each upstream `Task(...)` into `spawn_agent` -> `wait` -> `close_agent`.
+- Spawn only when the upstream workflow defines an agent role.
+- Use `.claude/agents/gsd-*.md` as role context for each spawned agent.
+- Do not advance workflow steps until wait and close complete.
+## Execution
+1. Parse "[phase] [--research] [--skip-research] [--gaps] [--skip-verify]" from the user input.
+2. Run init:
+node .claude/get-shit-done/bin/gsd-tools.js init plan-phase $PHASE --include state,roadmap,requirements,context,research,verification,uat --raw
+
+3. Load .claude/get-shit-done/workflows/plan-phase.md and execute it step-by-step.
+4. Translate each Task(...) in workflow into:
+   - spawn_agent with the matching role file context from .claude/agents/.
+   - wait for each spawned agent and apply returned output before moving forward.
+5. Preserve all gates and routing from upstream workflow.
+6. Preserve commit behavior using 
+node .claude/get-shit-done/bin/gsd-tools.js commit "message" --files ....
+7. If commit preflight fails (no git / no commit flag), proceed in read-only mode and report clearly.
+
+## Completion output
+- Summarize key artifacts created/updated and next recommended command.
+
+

package/.codex/prompts/gsd-progress.md

@@ -0,0 +1,43 @@
+---
+description: Check project progress, show context, and route to next action
+argument-hint: [none]
+---
+
+
+
+## Objective
+Check project progress, show context, and route to next action
+
+## Compatibility
+- Use .codex/skills/get-shit-done-codex semantics.
+- Treat upstream workflow as the source of truth.
+- Replace user-specific paths with workspace-relative paths (.claude/..., .planning/...).
+- Run engine commands through PowerShell: 
+node .claude/get-shit-done/bin/gsd-tools.js ...
+- Parse JSON with ConvertFrom-Json; parse key/value output when workflow uses KEY=value raw mode.
+- No jq / bash-only constructs.
+
+## Subagent lifecycle (required)
+
+- Translate each upstream `Task(...)` into `spawn_agent` -> `wait` -> `close_agent`.
+- Spawn only when the upstream workflow defines an agent role.
+- Use `.claude/agents/gsd-*.md` as role context for each spawned agent.
+- Do not advance workflow steps until wait and close complete.
+## Execution
+1. Parse [none] from the user input.
+2. Run init:
+node .claude/get-shit-done/bin/gsd-tools.js init progress --include state,roadmap,project,config --raw
+
+3. Load .claude/get-shit-done/workflows/progress.md and execute it step-by-step.
+4. Translate each Task(...) in workflow into:
+   - spawn_agent with the matching role file context from .claude/agents/.
+   - wait for each spawned agent and apply returned output before moving forward.
+5. Preserve all gates and routing from upstream workflow.
+6. Preserve commit behavior using 
+node .claude/get-shit-done/bin/gsd-tools.js commit "message" --files ....
+7. If commit preflight fails (no git / no commit flag), proceed in read-only mode and report clearly.
+
+## Completion output
+- Summarize key artifacts created/updated and next recommended command.
+
+

package/.codex/prompts/gsd-quick.md

@@ -0,0 +1,43 @@
+---
+description: Execute a quick task with GSD guarantees (atomic commits, state tracking) but skip optional agents
+argument-hint: [none]
+---
+
+
+
+## Objective
+Execute a quick task with GSD guarantees (atomic commits, state tracking) but skip optional agents
+
+## Compatibility
+- Use .codex/skills/get-shit-done-codex semantics.
+- Treat upstream workflow as the source of truth.
+- Replace user-specific paths with workspace-relative paths (.claude/..., .planning/...).
+- Run engine commands through PowerShell: 
+node .claude/get-shit-done/bin/gsd-tools.js ...
+- Parse JSON with ConvertFrom-Json; parse key/value output when workflow uses KEY=value raw mode.
+- No jq / bash-only constructs.
+
+## Subagent lifecycle (required)
+
+- Translate each upstream `Task(...)` into `spawn_agent` -> `wait` -> `close_agent`.
+- Spawn only when the upstream workflow defines an agent role.
+- Use `.claude/agents/gsd-*.md` as role context for each spawned agent.
+- Do not advance workflow steps until wait and close complete.
+## Execution
+1. Parse [none] from the user input.
+2. Run init:
+node .claude/get-shit-done/bin/gsd-tools.js init quick "$ARG" --raw
+
+3. Load .claude/get-shit-done/workflows/quick.md and execute it step-by-step.
+4. Translate each Task(...) in workflow into:
+   - spawn_agent with the matching role file context from .claude/agents/.
+   - wait for each spawned agent and apply returned output before moving forward.
+5. Preserve all gates and routing from upstream workflow.
+6. Preserve commit behavior using 
+node .claude/get-shit-done/bin/gsd-tools.js commit "message" --files ....
+7. If commit preflight fails (no git / no commit flag), proceed in read-only mode and report clearly.
+
+## Completion output
+- Summarize key artifacts created/updated and next recommended command.
+
+

package/.codex/prompts/gsd-remove-phase.md

@@ -0,0 +1,43 @@
+---
+description: Remove a future phase from roadmap and renumber subsequent phases
+argument-hint: <phase-number>
+---
+
+
+
+## Objective
+Remove a future phase from roadmap and renumber subsequent phases
+
+## Compatibility
+- Use .codex/skills/get-shit-done-codex semantics.
+- Treat upstream workflow as the source of truth.
+- Replace user-specific paths with workspace-relative paths (.claude/..., .planning/...).
+- Run engine commands through PowerShell: 
+node .claude/get-shit-done/bin/gsd-tools.js ...
+- Parse JSON with ConvertFrom-Json; parse key/value output when workflow uses KEY=value raw mode.
+- No jq / bash-only constructs.
+
+## Subagent lifecycle (required)
+
+- Translate each upstream `Task(...)` into `spawn_agent` -> `wait` -> `close_agent`.
+- Spawn only when the upstream workflow defines an agent role.
+- Use `.claude/agents/gsd-*.md` as role context for each spawned agent.
+- Do not advance workflow steps until wait and close complete.
+## Execution
+1. Parse <phase-number> from the user input.
+2. Run init:
+node .claude/get-shit-done/bin/gsd-tools.js init phase-op $ARG --raw
+
+3. Load .claude/get-shit-done/workflows/remove-phase.md and execute it step-by-step.
+4. Translate each Task(...) in workflow into:
+   - spawn_agent with the matching role file context from .claude/agents/.
+   - wait for each spawned agent and apply returned output before moving forward.
+5. Preserve all gates and routing from upstream workflow.
+6. Preserve commit behavior using 
+node .claude/get-shit-done/bin/gsd-tools.js commit "message" --files ....
+7. If commit preflight fails (no git / no commit flag), proceed in read-only mode and report clearly.
+
+## Completion output
+- Summarize key artifacts created/updated and next recommended command.
+
+

package/.codex/prompts/gsd-research-phase.md

@@ -0,0 +1,43 @@
+---
+description: Research how to implement a phase (standalone - usually use plan-phase instead)
+argument-hint: "[phase]"
+---
+
+
+
+## Objective
+Research how to implement a phase (standalone - usually use plan-phase instead)
+
+## Compatibility
+- Use .codex/skills/get-shit-done-codex semantics.
+- Treat upstream workflow as the source of truth.
+- Replace user-specific paths with workspace-relative paths (.claude/..., .planning/...).
+- Run engine commands through PowerShell: 
+node .claude/get-shit-done/bin/gsd-tools.js ...
+- Parse JSON with ConvertFrom-Json; parse key/value output when workflow uses KEY=value raw mode.
+- No jq / bash-only constructs.
+
+## Subagent lifecycle (required)
+
+- Translate each upstream `Task(...)` into `spawn_agent` -> `wait` -> `close_agent`.
+- Spawn only when the upstream workflow defines an agent role.
+- Use `.claude/agents/gsd-*.md` as role context for each spawned agent.
+- Do not advance workflow steps until wait and close complete.
+## Execution
+1. Parse "[phase]" from the user input.
+2. Run init:
+node .claude/get-shit-done/bin/gsd-tools.js init phase-op $ARG --raw
+
+3. Load .claude/get-shit-done/workflows/research-phase.md and execute it step-by-step.
+4. Translate each Task(...) in workflow into:
+   - spawn_agent with the matching role file context from .claude/agents/.
+   - wait for each spawned agent and apply returned output before moving forward.
+5. Preserve all gates and routing from upstream workflow.
+6. Preserve commit behavior using 
+node .claude/get-shit-done/bin/gsd-tools.js commit "message" --files ....
+7. If commit preflight fails (no git / no commit flag), proceed in read-only mode and report clearly.
+
+## Completion output
+- Summarize key artifacts created/updated and next recommended command.
+
+

package/.codex/prompts/gsd-resume-work.md

@@ -0,0 +1,43 @@
+---
+description: Resume work from previous session with full context restoration
+argument-hint: [none]
+---
+
+
+
+## Objective
+Resume work from previous session with full context restoration
+
+## Compatibility
+- Use .codex/skills/get-shit-done-codex semantics.
+- Treat upstream workflow as the source of truth.
+- Replace user-specific paths with workspace-relative paths (.claude/..., .planning/...).
+- Run engine commands through PowerShell: 
+node .claude/get-shit-done/bin/gsd-tools.js ...
+- Parse JSON with ConvertFrom-Json; parse key/value output when workflow uses KEY=value raw mode.
+- No jq / bash-only constructs.
+
+## Subagent lifecycle (required)
+
+- Translate each upstream `Task(...)` into `spawn_agent` -> `wait` -> `close_agent`.
+- Spawn only when the upstream workflow defines an agent role.
+- Use `.claude/agents/gsd-*.md` as role context for each spawned agent.
+- Do not advance workflow steps until wait and close complete.
+## Execution
+1. Parse [none] from the user input.
+2. Run init:
+node .claude/get-shit-done/bin/gsd-tools.js init resume --raw
+
+3. Load .claude/get-shit-done/workflows/resume-project.md and execute it step-by-step.
+4. Translate each Task(...) in workflow into:
+   - spawn_agent with the matching role file context from .claude/agents/.
+   - wait for each spawned agent and apply returned output before moving forward.
+5. Preserve all gates and routing from upstream workflow.
+6. Preserve commit behavior using 
+node .claude/get-shit-done/bin/gsd-tools.js commit "message" --files ....
+7. If commit preflight fails (no git / no commit flag), proceed in read-only mode and report clearly.
+
+## Completion output
+- Summarize key artifacts created/updated and next recommended command.
+
+

package/.codex/prompts/gsd-set-profile.md

@@ -0,0 +1,43 @@
+---
+description: Switch model profile for GSD agents (quality/balanced/budget)
+argument-hint: <profile>
+---
+
+
+
+## Objective
+Switch model profile for GSD agents (quality/balanced/budget)
+
+## Compatibility
+- Use .codex/skills/get-shit-done-codex semantics.
+- Treat upstream workflow as the source of truth.
+- Replace user-specific paths with workspace-relative paths (.claude/..., .planning/...).
+- Run engine commands through PowerShell: 
+node .claude/get-shit-done/bin/gsd-tools.js ...
+- Parse JSON with ConvertFrom-Json; parse key/value output when workflow uses KEY=value raw mode.
+- No jq / bash-only constructs.
+
+## Subagent lifecycle (required)
+
+- Translate each upstream `Task(...)` into `spawn_agent` -> `wait` -> `close_agent`.
+- Spawn only when the upstream workflow defines an agent role.
+- Use `.claude/agents/gsd-*.md` as role context for each spawned agent.
+- Do not advance workflow steps until wait and close complete.
+## Execution
+1. Parse <profile> from the user input.
+2. Run init:
+node .claude/get-shit-done/bin/gsd-tools.js state load --raw
+
+3. Load .claude/get-shit-done/workflows/set-profile.md and execute it step-by-step.
+4. Translate each Task(...) in workflow into:
+   - spawn_agent with the matching role file context from .claude/agents/.
+   - wait for each spawned agent and apply returned output before moving forward.
+5. Preserve all gates and routing from upstream workflow.
+6. Preserve commit behavior using 
+node .claude/get-shit-done/bin/gsd-tools.js commit "message" --files ....
+7. If commit preflight fails (no git / no commit flag), proceed in read-only mode and report clearly.
+
+## Completion output
+- Summarize key artifacts created/updated and next recommended command.
+
+

package/.codex/prompts/gsd-settings.md

@@ -0,0 +1,43 @@
+---
+description: Configure GSD workflow toggles and model profile
+argument-hint: [none]
+---
+
+
+
+## Objective
+Configure GSD workflow toggles and model profile
+
+## Compatibility
+- Use .codex/skills/get-shit-done-codex semantics.
+- Treat upstream workflow as the source of truth.
+- Replace user-specific paths with workspace-relative paths (.claude/..., .planning/...).
+- Run engine commands through PowerShell: 
+node .claude/get-shit-done/bin/gsd-tools.js ...
+- Parse JSON with ConvertFrom-Json; parse key/value output when workflow uses KEY=value raw mode.
+- No jq / bash-only constructs.
+
+## Subagent lifecycle (required)
+
+- Translate each upstream `Task(...)` into `spawn_agent` -> `wait` -> `close_agent`.
+- Spawn only when the upstream workflow defines an agent role.
+- Use `.claude/agents/gsd-*.md` as role context for each spawned agent.
+- Do not advance workflow steps until wait and close complete.
+## Execution
+1. Parse [none] from the user input.
+2. Run init:
+node .claude/get-shit-done/bin/gsd-tools.js config-ensure-section --raw
+
+3. Load .claude/get-shit-done/workflows/settings.md and execute it step-by-step.
+4. Translate each Task(...) in workflow into:
+   - spawn_agent with the matching role file context from .claude/agents/.
+   - wait for each spawned agent and apply returned output before moving forward.
+5. Preserve all gates and routing from upstream workflow.
+6. Preserve commit behavior using 
+node .claude/get-shit-done/bin/gsd-tools.js commit "message" --files ....
+7. If commit preflight fails (no git / no commit flag), proceed in read-only mode and report clearly.
+
+## Completion output
+- Summarize key artifacts created/updated and next recommended command.
+
+

package/.codex/prompts/gsd-update.md

@@ -0,0 +1,43 @@
+---
+description: Update GSD to latest version with changelog display
+argument-hint: [none]
+---
+
+
+
+## Objective
+Update GSD to latest version with changelog display
+
+## Compatibility
+- Use .codex/skills/get-shit-done-codex semantics.
+- Treat upstream workflow as the source of truth.
+- Replace user-specific paths with workspace-relative paths (.claude/..., .planning/...).
+- Run engine commands through PowerShell: 
+node .claude/get-shit-done/bin/gsd-tools.js ...
+- Parse JSON with ConvertFrom-Json; parse key/value output when workflow uses KEY=value raw mode.
+- No jq / bash-only constructs.
+
+## Subagent lifecycle (required)
+
+- Translate each upstream `Task(...)` into `spawn_agent` -> `wait` -> `close_agent`.
+- Spawn only when the upstream workflow defines an agent role.
+- Use `.claude/agents/gsd-*.md` as role context for each spawned agent.
+- Do not advance workflow steps until wait and close complete.
+## Execution
+1. Parse [none] from the user input.
+2. Run init:
+# No explicit init.
+
+3. Load .claude/get-shit-done/workflows/update.md and execute it step-by-step.
+4. Translate each Task(...) in workflow into:
+   - spawn_agent with the matching role file context from .claude/agents/.
+   - wait for each spawned agent and apply returned output before moving forward.
+5. Preserve all gates and routing from upstream workflow.
+6. Preserve commit behavior using 
+node .claude/get-shit-done/bin/gsd-tools.js commit "message" --files ....
+7. If commit preflight fails (no git / no commit flag), proceed in read-only mode and report clearly.
+
+## Completion output
+- Summarize key artifacts created/updated and next recommended command.
+
+

package/.codex/prompts/gsd-verify-work.md

@@ -0,0 +1,43 @@
+---
+description: Validate built features through conversational UAT
+argument-hint: "[phase number, for example 4]"
+---
+
+
+
+## Objective
+Validate built features through conversational UAT
+
+## Compatibility
+- Use .codex/skills/get-shit-done-codex semantics.
+- Treat upstream workflow as the source of truth.
+- Replace user-specific paths with workspace-relative paths (.claude/..., .planning/...).
+- Run engine commands through PowerShell: 
+node .claude/get-shit-done/bin/gsd-tools.js ...
+- Parse JSON with ConvertFrom-Json; parse key/value output when workflow uses KEY=value raw mode.
+- No jq / bash-only constructs.
+
+## Subagent lifecycle (required)
+
+- Translate each upstream `Task(...)` into `spawn_agent` -> `wait` -> `close_agent`.
+- Spawn only when the upstream workflow defines an agent role.
+- Use `.claude/agents/gsd-*.md` as role context for each spawned agent.
+- Do not advance workflow steps until wait and close complete.
+## Execution
+1. Parse "[phase number, for example 4]" from the user input.
+2. Run init:
+node .claude/get-shit-done/bin/gsd-tools.js init verify-work $PHASE --raw
+
+3. Load .claude/get-shit-done/workflows/verify-work.md and execute it step-by-step.
+4. Translate each Task(...) in workflow into:
+   - spawn_agent with the matching role file context from .claude/agents/.
+   - wait for each spawned agent and apply returned output before moving forward.
+5. Preserve all gates and routing from upstream workflow.
+6. Preserve commit behavior using 
+node .claude/get-shit-done/bin/gsd-tools.js commit "message" --files ....
+7. If commit preflight fails (no git / no commit flag), proceed in read-only mode and report clearly.
+
+## Completion output
+- Summarize key artifacts created/updated and next recommended command.
+
+

package/.codex/skills/get-shit-done-codex/SKILL.md

@@ -0,0 +1,65 @@
+---
+name: get-shit-done-codex
+description: Codex-compatible orchestration for get-shit-done workflows.
+metadata:
+  short-description: Translation layer from Claude GSD workflows to Codex
+---
+
+# Get-shit-done (GSD) Codex Skill
+
+Use this skill to run `gsd-*` prompts under `Codex` while preserving upstream GSD behavior.
+
+## Scope
+- Keep `.claude/get-shit-done` as the command engine and workflow source.
+- Keep `.claude/agents/gsd-*.md` as role contracts for subagents.
+- Keep `.codex/prompts/gsd-*.md` as Codex-native orchestrators.
+
+## Core rules
+1. Do not change command semantics.
+2. Do not change source workflow logic unless necessary for Codex command compatibility.
+3. Use workspace-relative paths.
+4. Preserve step ordering and gate behavior from workflow files.
+
+## Mandatory codex translations
+- Replace each `Task(...)` with: `spawn_agent` + `wait`.
+- Replace Bash/JQ patterns with PowerShell and `ConvertFrom-Json`.
+- Replace AskUserQuestion interactions with direct user prompts in chat.
+
+## Subagent lifecycle (required)
+- Spawn only when the upstream workflow explicitly defines a `Task(...)` role.
+- For each spawned subagent:
+  - Start with `spawn_agent` and provide the role source from `.claude/agents/gsd-*.md`.
+  - Wait for completion before proceeding: `wait`.
+  - Explicitly close the agent after it finishes (`close_agent`) to release resources.
+- Keep prompt-level sequencing intact; do not continue to subsequent steps until waiting + closure are complete.
+- If an agent fails or exceeds retry policy, stop that phase of workflow and surface a clear remediation step.
+
+## Required GSD roles (must map by name)
+- `gsd-project-researcher`
+- `gsd-research-synthesizer`
+- `gsd-roadmapper`
+- `gsd-phase-researcher`
+- `gsd-planner`
+- `gsd-plan-checker`
+- `gsd-executor`
+- `gsd-verifier`
+- `gsd-debugger`
+- `gsd-integration-checker`
+- `gsd-codebase-mapper`
+
+## Standard execution order
+1. Parse user input from `$ARGUMENTS`.
+2. Run required `gsd-tools init ...` if the workflow defines it.
+3. Read workflow path in `.claude/get-shit-done/workflows/`.
+4. Execute step-by-step, translating subagent calls.
+5. Preserve `commit_docs`, state updates, and file contracts.
+6. Report outcome and next action.
+
+## Commit discipline
+- Validate git first: `git rev-parse --is-inside-work-tree`.
+- Use `node .claude/get-shit-done/bin/gsd-tools.js commit ...` for all intended doc/execution writes.
+- If commit is not possible, continue execution with explicit warning and no commit.
+
+## Required references
+- `.codex/skills/get-shit-done-codex/references/compat.md`
+- `.codex/skills/get-shit-done-codex/references/windows.md`

package/.codex/skills/get-shit-done-codex/references/compat.md

@@ -0,0 +1,32 @@
+# Compatibility: Claude → Codex for get-shit-done
+
+## Tool mapping
+- `Task(...)` → `spawn_agent` + `wait` + `close_agent`.
+- `AskUserQuestion` → direct user interaction in chat.
+- Bash-style shell helpers → PowerShell commands.
+
+## Path mapping
+- `C:/Users/Ahmed/.claude/get-shit-done/...` → `.claude/get-shit-done/...`
+- `C:/Users/Ahmed/.claude/agents/...` → `.claude/agents/...`
+
+## CLI mapping
+- Use `node .claude/get-shit-done/bin/gsd-tools.js ...`.
+- Keep `--raw` when upstream workflow uses it.
+- Parse JSON by `ConvertFrom-Json`.
+
+## Subagent mapping
+- `subagent_type=gsd-*` maps to equivalent role contract in `.claude/agents/gsd-*.md`.
+- Unspecified `subagent_type` values default to command-context Codex agent behavior.
+
+## Required GSD subagents
+- `gsd-project-researcher`
+- `gsd-research-synthesizer`
+- `gsd-roadmapper`
+- `gsd-phase-researcher`
+- `gsd-planner`
+- `gsd-plan-checker`
+- `gsd-executor`
+- `gsd-verifier`
+- `gsd-debugger`
+- `gsd-integration-checker`
+- `gsd-codebase-mapper`

package/.codex/skills/get-shit-done-codex/references/windows.md

@@ -0,0 +1,23 @@
+# Windows + PowerShell conversion notes
+
+## JSON parse
+
+```powershell
+$raw = & node .claude/get-shit-done/bin/gsd-tools.js state load --raw
+$data = $raw | ConvertFrom-Json
+```
+
+## Key/value parse
+
+```powershell
+$raw = & node .claude/get-shit-done/bin/gsd-tools.js state load --raw
+$out = @{}
+$raw -split "`n" | ForEach-Object {
+  if ($_ -match '^(.*?)=(.*)$') { $out[$Matches[1]] = $Matches[2] }
+}
+```
+
+## Syntax notes
+- Replace jq with `ConvertFrom-Json`.
+- Avoid bash-only heredocs and command substitution.
+- Use quoted paths for PowerShell compatibility.