feat/mock cli

.github/workflows/cli-skill-review-reusable.yml

@@ -107,10 +107,6 @@ jobs:
           }
           EOF
 
-      - name: Sync cli-skill adapters
-        if: ${{ steps.cli_scope.outputs.proceed == 'true' }}
-        run: node scripts/sync-cli-skill-adapters.js
-
       - name: Install cli-skill for Pi discovery
         if: ${{ steps.cli_scope.outputs.proceed == 'true' }}
         run: |
@@ -167,47 +163,7 @@ jobs:
           COMMAND: ${{ inputs.command }}
         run: |
           set -euo pipefail
-
-          python3 - << 'PY'
-          import os
-          from pathlib import Path
-
-          marker = os.environ.get('MARKER', '<!-- cli-skill-report -->')
-          command = os.environ.get('COMMAND', '/cli-review')
-          success = os.environ.get('PI_SUCCESS', '')
-          cost = os.environ.get('PI_COST', '')
-          duration = os.environ.get('PI_DURATION', '')
-          input_tokens = os.environ.get('PI_INPUT_TOKENS', '')
-          output_tokens = os.environ.get('PI_OUTPUT_TOKENS', '')
-          run_url = os.environ.get('RUN_URL', '')
-          response = os.environ.get('PI_RESPONSE', '')
-
-          if not response.strip():
-              response = '_No response was returned by Pi._'
-
-          # Keep body safely below GitHub comment max size.
-          max_response_len = 50000
-          if len(response) > max_response_len:
-              response = response[:max_response_len] + "\n\n... _truncated_"
-
-          body = f"""{marker}
-          ## CLI Skill Report ({command})
-
-          | Metric | Value |
-          |---|---|
-          | Success | {success or 'unknown'} |
-          | Duration (s) | {duration or 'n/a'} |
-          | Cost (USD) | {cost or 'n/a'} |
-          | Tokens (in/out) | {input_tokens or 'n/a'} / {output_tokens or 'n/a'} |
-          | Workflow run | [view run]({run_url}) |
-
-          {response}
-          """
-
-          p = Path('pi-report-comment.md')
-          p.write_text(body.strip() + "\n", encoding='utf-8')
-          print(f"Wrote {p}")
-          PY
+          python3 scripts/build-pr-report.py
 
       - name: Fail when agent run is unsuccessful
         if: ${{ steps.cli_scope.outputs.proceed == 'true' && inputs.fail_on_agent_error && steps.pi.outputs.success != 'true' }}

cli-skill/commands/cli-review.md

@@ -16,8 +16,8 @@ CLI standard compliance review only.
 - Non-compliance and compliant evidence based on observed CLI behavior/docs/code
 - Use these rules to determine severity:
   * High: violations of command structure and naming, use of positional parameters, accessibility/color violations
-  * Medium: use if non-standard verbs for commands, inconsistent flag names, extremely high complexity (eg. created by >20 commands)
-  * Low: formatting violations
+  * Medium: use if non-standard verbs for commands, inconsistent flag names or usage, extremely high complexity (eg. created by >20 commands)
+  * Low: formatting violations, duplicate short/long flags
 
 ## Out Of Scope
 
@@ -43,7 +43,7 @@ Required sections:
 
 ## Summary Requirements
 The summary should list the number of violations, and their severity. It shall include these in a table. It shall give an overall score (Excellent = >95%, Very Good = >90%, Good = >80%, Room for Improvement = >60%, Need for Action = <=60%) 
-Start with a score of 100%, the number of commands N, and the weight of a single command W=1/N. For each High violation, reduce the score by 5*W, for each Medium violation by 2*W, and for each Small violation by 0.5*W.
+The score is calculated based on violations and set into relation with the size of the command set. Start with a score of 100%, the number of commands N, and the weight of a single command W=100/N. First, make a list of all the violations sorted by command. For each High violation, reduce the score by 5*W, for each Medium violation by 2*W, and for each Small violation by 0.5*W. USE THIS ALGORITHM, DO NOT REASON ABOUT IT, OR FIND ALTERNATIVE WAYS TO CALCULATE A SCORE.
 
 ## Compliance Matrix Requirements
 

knowledge_base/CLI.md

@@ -1 +1,4 @@
-This project has no CLI.
+# CLI scope
+
+This repository includes a mock task-management CLI fixture for standards-review testing.
+paths: [scripts/demo_cli.py]

scripts/build-pr-report.py

@@ -0,0 +1,45 @@
+#!/usr/bin/env python3
+import os
+from pathlib import Path
+
+
+def main() -> None:
+    marker = os.environ.get("MARKER", "<!-- cli-skill-report -->")
+    command = os.environ.get("COMMAND", "/cli-review")
+    success = os.environ.get("PI_SUCCESS", "")
+    cost = os.environ.get("PI_COST", "")
+    duration = os.environ.get("PI_DURATION", "")
+    input_tokens = os.environ.get("PI_INPUT_TOKENS", "")
+    output_tokens = os.environ.get("PI_OUTPUT_TOKENS", "")
+    run_url = os.environ.get("RUN_URL", "")
+    response = os.environ.get("PI_RESPONSE", "")
+
+    if not response.strip():
+        response = "_No response was returned by Pi._"
+
+    # Keep body safely below GitHub comment max size.
+    max_response_len = 50000
+    if len(response) > max_response_len:
+        response = response[:max_response_len] + "\n\n... _truncated_"
+
+    body = f"""{marker}
+## CLI Skill Report ({command})
+
+| Metric | Value |
+|---|---|
+| Success | {success or 'unknown'} |
+| Duration (s) | {duration or 'n/a'} |
+| Cost (USD) | {cost or 'n/a'} |
+| Tokens (in/out) | {input_tokens or 'n/a'} / {output_tokens or 'n/a'} |
+| Workflow run | [view run]({run_url}) |
+
+{response}
+"""
+
+    report_path = Path("pi-report-comment.md")
+    report_path.write_text(body.strip() + "\n", encoding="utf-8")
+    print(f"Wrote {report_path}")
+
+
+if __name__ == "__main__":
+    main()

scripts/demo_cli.py

@@ -0,0 +1,82 @@
+#!/usr/bin/env python3
+"""Intentionally non-compliant demo CLI for task-management review tests."""
+
+import argparse
+
+
+def build_parser() -> argparse.ArgumentParser:
+	parser = argparse.ArgumentParser(
+		prog="taskThing",
+		description="A wildly inconsistent task manager CLI that is intentionally bad for standards testing.",
+		epilog="Try basically anything. Some commands probably look right.",
+	)
+	parser.add_argument("-v", "--verbose", action="store_true", help="be extra talkative")
+	parser.add_argument("-q", "--quiet", action="store_true", help="say less maybe")
+	parser.add_argument("--format", "-f", choices=["txt", "json", "table"], default="txt")
+
+	commands = parser.add_subparsers(dest="what")
+
+	add = commands.add_parser(
+		"addTodo",
+		help="Add a todo thing into the todo area",
+		description="Create one todo item in a very flexible and not especially predictable way.",
+	)
+	add.add_argument("title", help="todo name")
+	add.add_argument("place", nargs="?", help="where it lives")
+	add.add_argument("owner", nargs="?", help="person maybe")
+	add.add_argument("-p", "--priority", dest="priority")
+	add.add_argument("--prio", dest="priority")
+	add.add_argument("-t", "--tag", action="append")
+	add.add_argument("--tags", help="comma separated tags")
+
+	remove = commands.add_parser(
+		"remove-todo-item",
+		help="Remove, delete, or otherwise get rid of a todo",
+		description="Deletes a task by whichever identifier feels useful at the time.",
+	)
+	remove.add_argument("todo")
+	remove.add_argument("extra", nargs="?", help="second positional just because")
+	remove.add_argument("-i", "--id")
+	remove.add_argument("--name")
+	remove.add_argument("-y", "--yes", action="store_true")
+	remove.add_argument("--force", "-F", action="store_true")
+
+	update = commands.add_parser(
+		"update",
+		help="Update something about todos",
+		description="A two-level command for changing todo-ish data in a not very disciplined way.",
+	)
+	update_commands = update.add_subparsers(dest="update_target")
+	update_todos = update_commands.add_parser(
+		"todos",
+		help="Update todo records in bulk or singular form",
+	)
+	update_todos.add_argument("todo")
+	update_todos.add_argument("field", nargs="?", help="what to change")
+	update_todos.add_argument("value", nargs="?", help="new value maybe")
+	update_todos.add_argument("-s", "--status")
+	update_todos.add_argument("--set-status")
+	update_todos.add_argument("--label")
+	update_todos.add_argument("--labels")
+
+	read = commands.add_parser(
+		"readTodos",
+		help="Read todo information in one of several inconsistent ways",
+		description="Shows one todo, many todos, or maybe all todos depending on the inputs you try.",
+	)
+	read.add_argument("todo", nargs="?", help="todo name or id or blank for everything")
+	read.add_argument("-a", "--all", action="store_true")
+	read.add_argument("--id")
+	read.add_argument("--sort")
+	read.add_argument("--order")
+	read.add_argument("--no-headers", action="store_true")
+
+	return parser
+
+
+def main() -> None:
+	build_parser().parse_args()
+
+
+if __name__ == "__main__":
+	main()