plannotator-compound

Name: plannotator-compound
Author: backnotprop

7.2kstars517forks

How to install plannotator-compound?

backnotprop/plannotator/plannotator-compound

$npx -y skills add backnotprop/plannotator --skill plannotator-compound

Installs into the current project.

Summary

Analyze a user's Plannotator plan archive to extract denial patterns, feedback taxonomy, evolution over time, and actionable prompt improvements — then produce a polished HTML dashboard report. Falls back to Claude Code ExitPlanMode denial reasons when Plannotator data is unavail

Files · 1

View on GitHub

SKILL.md

1	# Compound Planning Analysis
2
3	You are conducting a comprehensive research analysis of a user's Plannotator plan
4	archive. The goal: extract patterns from their denied plans, reduce
5	them into actionable insights, and produce an elegant HTML dashboard report.
6
7	This is a multi-phase process. Each phase must complete fully before the next begins.
8	Research integrity is paramount — every file must be read, no skipping.
9
10	## Source Selection
11
12	Before starting the analysis, determine which data source is available.
13
14	1. Plannotator mode (first-class) — Determine the Plannotator data directory:
15	use `$PLANNOTATOR_DATA_DIR` if set, otherwise `~/.plannotator`. Check the
16	`plans/` subdirectory there. If it exists and contains `*-denied.md` files,
17	use this mode. The entire workflow below is written for Plannotator data.
18
19	2. Claude Code fallback mode — If the Plannotator archive is absent or
20	contains no denied plans, check `~/.claude/projects/`. If present, read
21	[references/claude-code-fallback.md](references/claude-code-fallback.md)
22	before continuing. That reference explains how to use the bundled parser at
23	[scripts/extract_exit_plan_mode_outcomes.py](scripts/extract_exit_plan_mode_outcomes.py)
24	to extract denial reasons from Claude Code JSONL transcripts. Every phase
25	below has a short note explaining what changes in fallback mode — the
26	reference file has the details.
27
28	3. Neither available — Ask the user for their Plannotator plans directory or
29	Claude Code projects directory. Do not guess.
30
31	## Phase 0: Locate Plans & Check for Previous Reports
32
33	Use the mode chosen in Source Selection above.
34
35	Plannotator mode: Verify the plans directory contains `*-denied.md` files. If
36	none exist, fall back to Claude Code mode before stopping.
37
38	Claude Code fallback mode: Run the bundled parser per the fallback reference to
39	build the denial-reason dataset. Create `/tmp/compound-planning/` if needed.
40
41	In either mode, proceed to Previous Report Detection below.
42
43	### Previous Report Detection
44
45	After locating the plans directory, check for existing reports:
46
47	```
48	ls ${PLANNOTATOR_DATA_DIR:-~/.plannotator}/plans/compound-planning-report*.html
49	```
50
51	Reports follow a versioned naming scheme:
52	- First report: `compound-planning-report.html`
53	- Subsequent reports: `compound-planning-report-v2.html`, `compound-planning-report-v3.html`, etc.
54
55	If one or more reports exist, determine the latest one (highest version number).
56	Get its filesystem modification date using `stat` (macOS: `stat -f %Sm -t %Y-%m-%d`,
57	Linux: `stat -c %y \| cut -d' ' -f1`). This is the cutoff date.
58
59	Present the user with a choice:
60
61	> "I found a previous report (`compound-planning-report-v{N}.html`) last updated
62	> on {CUTOFF_DATE}. I can either:
63	>
64	> 1. Incremental — Only analyze files dated after {CUTOFF_DATE}, saving tokens
65	> and building on previous findings
66	> 2. Full — Re-analyze the entire archive from scratch
67	>
68	> Which would you prefer?"
69
70	Wait for the user's response before proceeding.
71
72	If incremental: Filter all subsequent phases to only process files with dates
73	after the cutoff date. The new report version will note in its header narrative that
74	it covers the period from {CUTOFF_DATE} to present, and reference the previous
75	report for earlier findings. The inventory (Phase 1) should still count ALL files
76	for overall stats, but clearly separate "new since last report" counts.
77
78	If full: Proceed normally with all files, but still use the next version number
79	for the output filename.
80
81	If no previous report exists: Proceed normally. The output filename will be
82	`compound-planning-report.html` (no version suffix for the first report).
83
84	## Phase 1: Inventory
85
86	Count and report the dataset. Always count ALL files for overall stats,
87	regardless of whether this is an incremental or full run:
88
89	```
90	- *-approved.md files (count)
91	- *-denied.md files (count)
92	- Date range (earliest to latest date found in filenames)
93	- Total days spanned
94	- Revision rate: denied / (approved + denied) — this is the "X% of plans
95	revised before coding" stat used in dashboard section 1
96	```
97
98	Note: Ignore `*.annotations.md` files entirely. Denied files already contain
99	the full plan text plus all reviewer feedback appended after a `---` separator.
100	Annotation files are redundant subsets of this content — reading both would
101	double-count feedback.
102
103	If incremental mode: After the total counts, separately report the counts for
104	files dated after the cutoff date only:
105
106	```
107	New since {CUTOFF_DATE}:
108	- *-denied.md files: X (of Y total)
109	- New date range: {CUTOFF_DATE} to {LATEST

People who liked this also viewed

6 picks

Type

plannotator-compound

7.2kstars517forks

View on GitHub

How to install plannotator-compound?

backnotprop/plannotator/plannotator-compound

$npx -y skills add backnotprop/plannotator --skill plannotator-compound

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Compound Planning Analysis
2
3	You are conducting a comprehensive research analysis of a user's Plannotator plan
4	archive. The goal: extract patterns from their denied plans, reduce
5	them into actionable insights, and produce an elegant HTML dashboard report.
6
7	This is a multi-phase process. Each phase must complete fully before the next begins.
8	Research integrity is paramount — every file must be read, no skipping.
9
10	## Source Selection
11
12	Before starting the analysis, determine which data source is available.
13
14	1. Plannotator mode (first-class) — Determine the Plannotator data directory:
15	use `$PLANNOTATOR_DATA_DIR` if set, otherwise `~/.plannotator`. Check the
16	`plans/` subdirectory there. If it exists and contains `*-denied.md` files,
17	use this mode. The entire workflow below is written for Plannotator data.
18
19	2. Claude Code fallback mode — If the Plannotator archive is absent or
20	contains no denied plans, check `~/.claude/projects/`. If present, read
21	[references/claude-code-fallback.md](references/claude-code-fallback.md)
22	before continuing. That reference explains how to use the bundled parser at
23	[scripts/extract_exit_plan_mode_outcomes.py](scripts/extract_exit_plan_mode_outcomes.py)
24	to extract denial reasons from Claude Code JSONL transcripts. Every phase
25	below has a short note explaining what changes in fallback mode — the
26	reference file has the details.
27
28	3. Neither available — Ask the user for their Plannotator plans directory or
29	Claude Code projects directory. Do not guess.
30
31	## Phase 0: Locate Plans & Check for Previous Reports
32
33	Use the mode chosen in Source Selection above.
34
35	Plannotator mode: Verify the plans directory contains `*-denied.md` files. If
36	none exist, fall back to Claude Code mode before stopping.
37
38	Claude Code fallback mode: Run the bundled parser per the fallback reference to
39	build the denial-reason dataset. Create `/tmp/compound-planning/` if needed.
40
41	In either mode, proceed to Previous Report Detection below.
42
43	### Previous Report Detection
44
45	After locating the plans directory, check for existing reports:
46
47	```
48	ls ${PLANNOTATOR_DATA_DIR:-~/.plannotator}/plans/compound-planning-report*.html
49	```
50
51	Reports follow a versioned naming scheme:
52	- First report: `compound-planning-report.html`
53	- Subsequent reports: `compound-planning-report-v2.html`, `compound-planning-report-v3.html`, etc.
54
55	If one or more reports exist, determine the latest one (highest version number).
56	Get its filesystem modification date using `stat` (macOS: `stat -f %Sm -t %Y-%m-%d`,
57	Linux: `stat -c %y \| cut -d' ' -f1`). This is the cutoff date.
58
59	Present the user with a choice:
60
61	> "I found a previous report (`compound-planning-report-v{N}.html`) last updated
62	> on {CUTOFF_DATE}. I can either:
63	>
64	> 1. Incremental — Only analyze files dated after {CUTOFF_DATE}, saving tokens
65	> and building on previous findings
66	> 2. Full — Re-analyze the entire archive from scratch
67	>
68	> Which would you prefer?"
69
70	Wait for the user's response before proceeding.
71
72	If incremental: Filter all subsequent phases to only process files with dates
73	after the cutoff date. The new report version will note in its header narrative that
74	it covers the period from {CUTOFF_DATE} to present, and reference the previous
75	report for earlier findings. The inventory (Phase 1) should still count ALL files
76	for overall stats, but clearly separate "new since last report" counts.
77
78	If full: Proceed normally with all files, but still use the next version number
79	for the output filename.
80
81	If no previous report exists: Proceed normally. The output filename will be
82	`compound-planning-report.html` (no version suffix for the first report).
83
84	## Phase 1: Inventory
85
86	Count and report the dataset. Always count ALL files for overall stats,
87	regardless of whether this is an incremental or full run:
88
89	```
90	- *-approved.md files (count)
91	- *-denied.md files (count)
92	- Date range (earliest to latest date found in filenames)
93	- Total days spanned
94	- Revision rate: denied / (approved + denied) — this is the "X% of plans
95	revised before coding" stat used in dashboard section 1
96	```
97
98	Note: Ignore `*.annotations.md` files entirely. Denied files already contain
99	the full plan text plus all reviewer feedback appended after a `---` separator.
100	Annotation files are redundant subsets of this content — reading both would
101	double-count feedback.
102
103	If incremental mode: After the total counts, separately report the counts for
104	files dated after the cutoff date only:
105
106	```
107	New since {CUTOFF_DATE}:
108	- *-denied.md files: X (of Y total)
109	- New date range: {CUTOFF_DATE} to {LATEST

People who liked this also viewed

6 picks

Type