@@ -11,6 +11,21 @@ Humanize creates an iterative feedback loop with two phases:
1111
1212The loop continues until all acceptance criteria are met or no issues remain.
1313
14+ ## Typical Planning Flow
15+
16+ 1 . Generate the initial implementation plan:
17+ ``` bash
18+ /humanize:gen-plan --input draft.md --output docs/plan.md
19+ ```
20+ 2 . If the plan is reviewed with ` CMT: ` ... ` ENDCMT ` annotations, refine it and generate a QA ledger:
21+ ``` bash
22+ /humanize:refine-plan --input docs/plan.md
23+ ```
24+ 3 . Start the RLCR loop on the refined plan:
25+ ``` bash
26+ /humanize:start-rlcr-loop docs/plan.md
27+ ```
28+
1429## Commands
1530
1631| Command | Purpose |
@@ -68,6 +83,7 @@ OPTIONS:
6883 --discussion Use discussion mode (iterative Claude/Codex convergence rounds)
6984 --direct Use direct mode (skip convergence rounds, proceed immediately to plan)
7085 -h, --help Show help message
86+ ```
7187
7288The gen-plan command transforms rough draft documents into structured implementation plans.
7389
@@ -78,7 +94,9 @@ Workflow:
78944 . Engages user to resolve any issues found
79955 . Generates a structured plan.md with acceptance criteria
80966 . Optionally starts ` /humanize:start-rlcr-loop ` if ` --auto-start-rlcr-if-converged ` conditions are met
81- ```
97+
98+ If reviewers later annotate the generated plan with ` CMT: ` ... ` ENDCMT ` blocks, run
99+ ` /humanize:refine-plan --input <plan.md> ` before starting or resuming implementation.
82100
83101### refine-plan
84102
0 commit comments