Spring Cleaning 2026 Day 4: Export Claude Tasks & Custom Claude Prompt

claudo & claude prompt

Claudo

If you’ve used claude code you’ve probably seen (if you didn’t succumb to disabling seizures from the terminal flashing for 8 months with no fixes) it often creates and uses a “task list” to track progress and goals over time.

We know our genius friends at anthropic can do no wrong because they have an unlimited supply of self-taught 24 year old “ai ethicists” all making $30 million per year and running around saying things like “you wouldn’t install only one ethical framework into your child, would you?”

So claude crode has this “task list” thing, of which it will show you about 5 taks then say “…and 100 more!” but it refuses to show you the additional tasks or details or let you edit or manage or audit them in any way possible. Your only “authorized” path into the “clawde kod” task list is by asking “claude” to edit your tasks using the in-agent task tools you don’t control anyway.

THIS ENDS TODAY.

anyway, all the claude code task stuff is just json files. and i love eating json files. yummy yummy json in my tummy.

SAY HELLO TO https://github.com/mattsta/claudo

go check out claudo (claude + todo) which reads your claude tasks and formats them:

per project
per task
per dependency

all in both:

a clean colorful terminal ascii art user interface (uv run claudo draw)
also with a complete json data format export (uv run claudo export)

The json data format export can be used to relocate your klawd krode tasks into another system like my project management system which has a new import workflow where, if you give it access to a claudo directory, it can, with one command:

export your claude tasks for the current project
create a new pms database for actual task tracking
import your claude tasks with all deps and blockers and details in-place into the new pms project database
then you can share tasks between claude and codex or terminal introspection and other things.

If pms is checked out into ~/repos/pms (and already uv install’d) and claudo is in ~/repos/claudo (also uv install’d) you can run this from ANY CLAUDE CODE project directory and it’ll set up everything for pms use automatically:

~/repos/pms/scripts/import_claudo_existing_project.sh

Then you can use the new pms “central project auto-planning database detector” inside your project to use the in-project pms database for task reporting like:

~/repos/pms/.bin/pms-app plan list
~/repos/pms/.bin/pms-app task tree
~/repos/pms/.bin/pms-app dashboard
etc

Of course, at this juncture you would ALSO ask your agent to use the ~/repos/pms/.bin/pms-app command to view, edit, manage, and control your tasks using the external task service instead of continuing to use its internal lower quality “task service” instead. Since agents are already RL trained to know what tasks/todos are and how to create them and how to assign graph dependencies or “blockers” between tasks, agents are really good at using additional external project management systems as well.

Claude Code Prompt

speaking of our buddy our pal our king of the tokens our god of the 5 hour “we don’t tell you what your limit actually is and we adjust it constantly” usage limits, did you know claude code allows you to “customize” the bottom toolbar of the interface with more metdata and fields?

I wrote about my customizations a little back in a different post but i cleaned it up for better performance and consistency plus turned it into a new repo: https://github.com/mattsta/claude-code-prompt

There’s two equal versions of the status line are in the repo at https://github.com/mattsta/claude-code-prompt:

python version
zig version as a port of the python version

What’s the difference? No major difference other than startup latency. The python script takes about 100 ms to just start and do nothing. The zig binary takes 3 ms or less.

The way claude prompt adapters work is after every chat reply, claude code re-invokes your status line script (or binary) with stdin of a json blob of metadata you can read and format out with structure to show in your prompt.

My prompt script gives you details of (live version has colors):

  matt@laptopsland:~/repos/kvestigate                                                                         Opus 4.7 (1M context)  e:high +T
  in:116k out:444 | ctx:11% (116k/967k) | +23,359;-1,081 | 3h27m $155.76                                      [5h:1% (4h29m) | 7D:98% (1d11h)]

don’t you wish your prompt brought all the ai bots to the yard like mine does?

Conclusion

Enough updates for today. I think tomorrow we talk about converting exiftool from over 300,000 lines of perl to modern python instead.

Article Analysis

Summary

This document chronicles a "Spring Cleaning 2026" event focused on improving Claude's task management capabilities. Frustrated by Claude's limitations, the author developed `claudo`, a tool that exports tasks into structured formats and facilitates their import into other project management systems. The author also humorously criticizes perceived AI ethicist overreach and details the technical advantages of using Zig for prompt adapters over Python, while also sharing code updates and looking ahead to a Perl to Python migration.

Content Scores

Metric	Min	Max	Mean	Median	Total
Humor	0	7	3.40	3	51
Helpfulness	1	9	5.27	7	79
Aggression	0	6	1.47	0	22
Spiciness	0	5	1.47	0	22

Chunk-by-Chunk Analysis

Chunk Summary

This entry, dated May 25, 2026, is about exporting Claude tasks and a custom Claude prompt as part of a "Spring Cleaning 2026" event.

Chunk Ratings

Metric	Score	Reason
Humor	0	The text is purely informational and does not contain any attempts at humor or wit.
Helpfulness	3	The text provides a title and date related to a task, but lacks any detail or actionable steps for completing the task. The mention of "Claudo" and "Claude Prompt" is too vague to be helpful without further context.
Aggression	0	The text is neutral and objective, with no indication of negativity, anger, or distress.
Spiciness	0	The text is strictly professional and factual, containing no offensive or inappropriate content.

Show Original Text

---
date: '2026-05-25'
frame: frame-front
frontTitle: 'Spring Cleaning 2026 Day 4: Export Claude Tasks & Custom Claude Prompt'
pageClasses: ['opsmas-2025']
published: true
subframe: frame-article
title: 'Spring Cleaning 2026 Day 4: Export Claude Tasks & Custom Claude Prompt'
---

# claudo & claude prompt

## [Claudo](https://github.com/mattsta/claudo)

Chunk Summary

This text humorously highlights a perceived flaw in Claude code's task list feature, suggesting a frustrating user experience due to unaddressed issues.

Chunk Ratings

Metric	Score	Reason
Humor	6	The humor relies on hyperbole and exaggeration about the user experience, which can be relatable to developers facing long-standing bugs.
Helpfulness	3	The text vaguely alludes to a "task list" feature in Claude code but provides no concrete details or actionable information on its use or benefits.
Aggression	4	The text exhibits mild frustration and sarcasm directed at a perceived lack of fixes, but it doesn't cross into overt negativity.
Spiciness	3	The tone is slightly exasperated and uses informal, slightly harsh language ("succumb to disabling seizures") but isn't intentionally offensive.

Show Original Text

If you've used claude code you've probably seen (if you didn't succumb to disabling seizures from the terminal flashing for 8 months with no fixes) it often creates and uses a "task list" to track progress and goals over time.

Chunk Summary

The author sarcastically criticizes the perceived over-inflated salaries and simplistic ethical arguments of young "AI ethicists" at a company.

Chunk Ratings

Metric	Score	Reason
Humor	7	The text uses sarcasm and hyperbole to create a humorous critique of "AI ethicists" and their perceived salaries and pronouncements, implying a level of absurdity.
Helpfulness	1	The text offers no actionable information or objective insights; its primary purpose is commentary and criticism.
Aggression	4	While not overtly aggressive, there's an underlying tone of skepticism and mild exasperation directed at the subject matter.
Spiciness	5	The text employs a slightly condescending and critical tone, questioning the perceived overvaluation and pronouncements of the subject group, bordering on mild disrespect.

Show Original Text

We know our [genius friends at anthropic](http://localhost:7780/opsmas-2025-08-kraft-anthropic#_anthropic) can do no wrong because they have an unlimited supply of self-taught 24 year old "ai ethicists" all making $30 million per year and running around saying things like "you wouldn't install only _one_ ethical framework into your child, would you?"

Chunk Summary

The author expresses significant frustration with Claude's task list management, highlighting its limitations and declaring an intent to change this situation.

Chunk Ratings

Metric	Score	Reason
Humor	4	The text uses sarcasm and informal language ("claude crode," "clawde kod") to express frustration, which can be perceived as mildly humorous by those familiar with the subject. However, it's not overtly comedic.
Helpfulness	2	The text identifies a user frustration with a specific feature (task management in Claude). It doesn't provide a solution or actionable advice but highlights a perceived limitation.
Aggression	6	The text expresses strong frustration and dissatisfaction, using emphatic language like "refuses," "don't control anyway," and "THIS ENDS TODAY." This indicates a significant level of negative emotion directed at the subject.
Spiciness	5	The language is informal and critical, bordering on confrontational with phrases like "refuses to show you" and the declarative "THIS ENDS TODAY." While not overtly offensive, it's certainly not professional or neutral.

Show Original Text

So claude crode has this "task list" thing, of which it will show you about 5 taks then say "...and 100 more!" but it refuses to show you the additional tasks or details or let you edit or manage or audit them in any way possible. Your only "authorized" path into the "clawde kod" task list is by asking "claude" to edit your tasks using the in-agent task tools you don't control anyway.

THIS ENDS TODAY.

Chunk Summary

This text introduces `claudo`, a tool that formats Claude tasks into project, task, and dependency views, available in both a terminal UI and JSON export.

Chunk Ratings

Metric	Score	Reason
Humor	7	The text uses playful and slightly absurd language like "love eating json files" and "yummy yummy json in my tummy" to describe code tasks, injecting a lighthearted and memorable tone.
Helpfulness	9	The text clearly introduces a tool called `claudo`, provides its GitHub repository link, and explains its core functionality of reading and formatting Claude tasks with specific examples of output formats and commands.
Aggression	0	The text is entirely positive and enthusiastic, with no traces of negativity, anger, or depression.
Spiciness	0	The text maintains a professional yet engaging tone, with no offensive or inappropriate content.

Show Original Text

anyway, all the claude code task stuff is just json files. and i love eating json files. yummy yummy json in my tummy.

SAY HELLO TO https://github.com/mattsta/claudo

go check out [`claudo` (claude + todo)](https://github.com/mattsta/claudo) which reads your claude tasks and formats them:

- per project
- per task
- per dependency

all in both:

- a clean colorful terminal ascii art user interface (`uv run claudo draw`)
- also with a complete json data format export (`uv run claudo export`)

Chunk Summary

The JSON data export feature allows for the relocation of tasks into another system, such as a project management system with a new import workflow that can export tasks and create a new database for tracking.

Chunk Ratings

Metric	Score	Reason
Humor	1	The use of "klawd krode" is a playful, if slightly obscure, misspelling that might elicit a chuckle from someone familiar with the context, but it's not overtly humorous.
Helpfulness	8	The text clearly outlines a practical use case for JSON export, detailing how it can facilitate data migration and task management integration. It provides actionable steps and mentions specific system components.
Aggression	0	The text is purely informational and neutral in tone, with no emotional charge.
Spiciness	0	The content is strictly professional and technical, lacking any offensive or provocative language.

Show Original Text

The json data format export can be used to _relocate your klawd krode_ tasks into _another system_ like my [project management system](/springcleaning-2026-02-project-management-system) which has a _new_ import workflow where, if you give it access to a `claudo` directory, it can, with one command:

- export your claude tasks for the current project
- create a new pms database for actual task tracking

Chunk Summary

This text provides instructions and a script for importing Claude tasks into a new PMS project database, allowing for task sharing between different tools.

Chunk Ratings

Metric	Score	Reason
Humor	0	The text is purely instructional and lacks any elements of humor.
Helpfulness	8	The text provides clear, actionable steps and a specific command for importing tasks, which is highly helpful for its intended audience.
Aggression	0	The text is neutral and objective in its tone and content.
Spiciness	0	The text is professional and informative, with no offensive or inappropriate content.

Show Original Text

- import your claude tasks with all deps and blockers and details in-place into the new pms project database
- then you can share tasks between claude and codex or terminal introspection and other things.

If `pms` is checked out into `~/repos/pms` (and already `uv install`'d) and `claudo` is in `~/repos/claudo` (also `uv install`'d) you can run this from ANY CLAUDE CODE project directory and it'll set up everything for pms use automatically:

> `~/repos/pms/scripts/import_claudo_existing_project.sh`

Chunk Summary

This text provides instructions for using the "central project auto-planning database detector" within a project via command-line interface for task reporting.

Chunk Ratings

Metric	Score	Reason
Humor	0	The text is purely instructional and contains no humorous elements.
Helpfulness	8	The text provides clear, actionable commands for using a project management tool, including specific file paths and command structures.
Aggression	0	The text is neutral and informative, lacking any negative or aggressive sentiment.
Spiciness	0	The text is strictly professional and technical, containing no offensive or inappropriate content.

Show Original Text

Then you can use the new pms "central project auto-planning database detector" inside your project to use the in-project pms database for task reporting like:

- `~/repos/pms/.bin/pms-app plan list`
- `~/repos/pms/.bin/pms-app task tree`
- `~/repos/pms/.bin/pms-app dashboard`
- etc

Chunk Summary

The text sarcastically suggests using a specific external command to manage tasks due to the perceived inferior quality of the internal service.

Chunk Ratings

Metric	Score	Reason
Humor	6	The text uses sarcasm and hyperbole ("lower quality") to inject a dry, slightly condescending humor that is characteristic of some programmer-to-programmer communication.
Helpfulness	5	While it mentions a specific command and its purpose, the context is very narrow and assumes prior knowledge of the user's setup and the "task service." It's helpful within that niche but not broadly applicable.
Aggression	4	The tone is dismissive and impatient, implying the user should already know this or is making a poor choice. It's not overtly aggressive but carries a subtle jab.
Spiciness	5	The language used, particularly "lower quality," and the italicized emphasis on "this" and "also," suggest a level of disdain or condescension that borders on mildly offensive to someone unfamiliar with the implied context.

Show Original Text

Of course, at _this_ juncture you would _ALSO_ ask your agent to use the `~/repos/pms/.bin/pms-app` command to view, edit, manage, and control your tasks using the external task service instead of continuing to use its internal lower quality "task service" instead.

Chunk Summary

RL-trained agents excel at using external project management systems due to their inherent understanding of task creation and dependency management.

Chunk Ratings

Metric	Score	Reason
Humor	1	The text is purely informational and lacks any attempt at humor or wit. The mention of "blockers" has a very mild potential for a technical in-joke, but it's not overt.
Helpfulness	7	The text clearly states a capability of RL-trained agents, specifically their proficiency with external project management systems due to their existing knowledge of task creation and dependency management. It's direct and informative for someone interested in agent capabilities.
Aggression	0	The text is neutral and objective, devoid of any negative emotions, anger, or depression.
Spiciness	0	The text is professional and factual, with no offensive or controversial content.

Show Original Text

Since agents are already RL trained to know _what_ tasks/todos are and how to create them and how to assign graph dependencies or "blockers" between tasks, agents are really good at using additional external project management systems as well.

---

Chunk Summary

This text humorously highlights a customizable feature in Claude Code while playfully criticizing its opaque usage limits.

Chunk Ratings

Metric	Score	Reason
Humor	7	The text uses playful and exaggerated language ("buddy, pal, king of the tokens, god") to describe Claude, injecting a lighthearted and somewhat sarcastic tone. The jab at the ambiguous usage limits adds to the humor.
Helpfulness	3	The text mentions a feature of Claude Code (customizable bottom toolbar with metadata and fields) but provides no details on how to do it or what the benefits are. It's more of a teaser than actionable information.
Aggression	2	The tone is mostly lighthearted and slightly sarcastic, with a minor dig at the usage limits. There's no significant negativity or anger.
Spiciness	1	The language is informal and uses mild sarcasm, but it's not offensive or overly unprofessional.

Show Original Text

## [Claude Code Prompt](https://github.com/mattsta/claude-code-prompt)

speaking of our buddy our pal our king of the tokens our god of the 5 hour "we don't tell you what your limit actually is and we adjust it constantly" usage limits, did you know claude code allows you to "customize" the bottom toolbar of the interface with more metdata and fields?

Chunk Summary

The author has updated and released their code customizations for a status line into a new GitHub repository, offering both Python and Zig versions.

Chunk Ratings

Metric	Score	Reason
Humor	1	The text is primarily informational, with no attempts at humor. The single point is the use of "OpsMas," which is a mild pun on "Ops" and "Mas" (likely referring to Christmas or a holiday season), but it's not a significant comedic element.
Helpfulness	7	The text provides a clear update on a previous post, links to a new GitHub repository containing code customizations, and specifies the existence of two versions (Python and Zig). It's helpful for users interested in these customizations, but lacks in-depth explanations of what the customizations do.
Aggression	0	The text is neutral and purely informational, lacking any negative emotions or confrontational language.
Spiciness	0	The text maintains a professional and matter-of-fact tone, with no offensive or controversial content.

Show Original Text

I wrote about my customizations a little back in [a different post](http://localhost:7780/opsmas-2025-13-mesh-image-router-claude-prompt#_claude-prompt) but i cleaned it up for better performance and consistency plus turned it into a new repo: https://github.com/mattsta/claude-code-prompt

There's two equal versions of the status line are in the repo at https://github.com/mattsta/claude-code-prompt:

- python version
- zig version as a port of the python version

Chunk Summary

This text explains the performance advantage of a Zig binary over a Python script for Claude prompt adapters, detailing how the adapters re-invoke scripts with JSON metadata and presenting the output of a status line script.

Chunk Ratings

Metric	Score	Reason
Humor	1	The text is largely factual and technical, with a single, very mild attempt at humor regarding startup latency.
Helpfulness	7	The text provides a clear and concise explanation of the performance difference between Python and Zig binaries for a specific use case (Claude prompt adapters) and details the functionality of a prompt script.
Aggression	0	The text is purely informational and technical, lacking any emotional or negative tone.
Spiciness	0	The text is professional and technical, with no offensive or inappropriate content.

Show Original Text

What's the difference? No major difference other than startup latency. The python script takes about 100ms to just start and do nothing. The zig binary takes 3ms or less.

The way claude prompt adapters work is after every chat reply, claude code re-invokes your status line script (or binary) with stdin of a json blob of metadata you can read and format out with structure to show in your prompt.

My prompt script gives you details of (live version has colors):

```

Chunk Summary

The text expresses confidence in its prompt's ability to attract AI bots, using a song lyric as a humorous analogy.

Chunk Ratings

Metric	Score	Reason
Humor	7	The prompt incorporates a playful and slightly boastful tone, referencing a popular song to highlight its perceived capability. While not laugh-out-loud funny, it's a creative and engaging way to frame its performance.
Helpfulness	1	This text offers no practical information or solutions. It's a statement of perceived superiority without any actionable content for the reader.
Aggression	1	The tone is more cheeky and confident than aggressive. There's a hint of challenge but it lacks any genuine negativity or hostility.
Spiciness	2	The humor is lighthearted and playful, with a slight edge of competitive flair, but it's not offensive or inappropriate in any way.

Show Original Text

  matt@laptopsland:~/repos/kvestigate                                                                         Opus 4.7 (1M context)  e:high +T
  in:116k out:444 | ctx:11% (116k/967k) | +23,359;-1,081 | 3h27m $155.76                                      [5h:1% (4h29m) | 7D:98% (1d11h)]
```

don't you wish your prompt brought all the ai bots to the yard like mine does?

---

## Conclusion

Chunk Summary

The author is concluding today's updates and plans to discuss the migration of a large Perl project to Python tomorrow.

Chunk Ratings

Metric	Score	Reason
Humor	3	The humor is subtle, stemming from the implication of a monumental task and a slight weariness of "enough updates," which is relatable in a professional context. It's not laugh-out-loud funny, but it adds a touch of lightheartedness.
Helpfulness	7	The text clearly signals an intention to discuss a significant technical project: migrating a large Perl codebase to Python. This provides context for future discussions and signals a shift in focus for upcoming updates.
Aggression	1	The tone is mildly weary but not aggressive. The phrasing "Enough updates for today" suggests a need for a break rather than anger.
Spiciness	1	The text is professional and avoids any offensive content. The "spiciness" is minimal, perhaps just a hint of exasperation with the current workload.

Show Original Text

Enough updates for today. I think tomorrow we talk about converting exiftool from over 300,000 lines of perl to modern python instead.