SkillAgentSearch skills...

os-big-picture

ALWAYS invoke this skill when the user asks where the project as a whole stands - "where are we", "what's the big picture", "what have we built", "what is in this project", "map the project", "what does this thing even do", "what is stale", "what can we delete", "update the roadmap" - in any languag…

Install / Use

npx skills add kharmanskyi/open-steps --skill os-big-picture

Installs into whichever agent you are using.

About this skill
📄

SKILL.md

Installable skill definition

Quality Score

89/100

Supported Platforms

Universal

Our assessment of os-big-picture

os-big-picture scores 89/100 on our quality scale, 30th of 60 Product Management skills we index (top 50%).

Its SKILL.md is 7.5 KB long, well organised into 13 sections with 3 code examples: a thorough specification that gives an agent plenty to work with.

It has 902 GitHub stars, a meaningful sign that others use it.

Substance
29/30
Structure
18/20
Description
15/15
Adoption
13/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated about 2 months ago, so os-big-picture is actively maintained.
  • Our last check on 2026-10-31 found the source still online.
  • It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
  • Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-10-04. Automated pattern scan on 2026-10-04. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

os-big-picture compared with similar skills

All 4 of these similar skills score higher than os-big-picture; compare them before choosing.

SkillScoreStarsUpdatedFormat
os-big-picture (this skill)by kharmanskyi899022mo agoSKILL.md
Agent-Reachby Panniantong100112.3k8d agoCLAUDE.md
x402by browser-use100121.0k3d agoSKILL.md
algorithmic-artby anthropics100177.9k55d agoSKILL.md
pptxby anthropics100177.9k55d agoSKILL.md

Frequently asked questions

How do I install os-big-picture?
Run npx skills add kharmanskyi/open-steps --skill os-big-picture. The install tabs above show the steps for each supported agent.
Which AI agents does os-big-picture work with?
It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
Is os-big-picture safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is os-big-picture still maintained?
The repository was last updated about 2 months ago, so os-big-picture is actively maintained.

name: os-big-picture description: >- ALWAYS invoke this skill when the user asks where the project as a whole stands - "where are we", "what's the big picture", "what have we built", "what is in this project", "map the project", "what does this thing even do", "what is stale", "what can we delete", "update the roadmap" - in any language, and whenever a session report has just been written. It keeps one file, BIG-PICTURE.md: what the product is, every feature with how far it got, which parts nobody has touched, and what is queued next. Age and wiring are measured from git; how far a feature got comes from the reports or says "not checked". Where a tracker is connected it offers to open the queue as tickets, after a yes and never before. Never invents work, never deletes code. allowed-tools:

  • "Read(~/.claude/open-steps/**)"
  • "Edit(BIG-PICTURE.md)"
  • "Write(BIG-PICTURE.md)"
  • "Bash(bash ${CLAUDE_SKILL_DIR}/scripts/census.sh *)"
  • "Bash(git rev-parse --git-dir)"
  • "Bash(gh repo view *)"
  • "Bash(gh issue list *)"
  • "Bash(gh issue create *)"

os-big-picture

One file, BIG-PICTURE.md, where os-whats-next already looks for a backlog. It answers "where are we now" - not a plan, nothing in it a promise. Write it in the user's language; paths stay English.

When to use it. The user asks where the project stands, or there is no BIG-PICTURE.md yet → the full pass: describe, census, signals, backlog. A session report was just written → the fold-in: only the rows that session touched. Both re-measure, because a session that moves no row is the one after which the dates go quietly stale.

Step 1 - find the file, and never clobber it

BIG-PICTURE.md in the project root, and it may already be somebody's own work. Everything this skill writes lives between <!-- open-steps:begin --> and <!-- open-steps:end -->; no markers → append at the end; no file → create it, and add it to this clone's ignore list, never to .gitignore:

f=BIG-PICTURE.md; e="$(git rev-parse --git-dir)/info/exclude"
git check-ignore -q "$f" || git ls-files --error-unmatch "$f" >/dev/null 2>&1 \
  || grep -qxF "$f" "$e" 2>/dev/null || printf '%s\n' "$f" >> "$e"

Each of the three is a reason to leave it alone. Say in one line that you did it; git add -f BIG-PICTURE.md shares it instead. Never read Last worked on or Signal back out of the file - only Stage and the queue carry forward.

Step 2 - the census, measured by the script

bash ${CLAUDE_SKILL_DIR}/scripts/census.sh .

Where your tool does not set ${CLAUDE_SKILL_DIR}, run it from the project root as bash <this skill's folder>/scripts/census.sh .. If it cannot run at all, say so and mark every measured column "not checked"; never estimate them.

An AGE line, then one measured PART row per part: path, last worked on, commits in six months, whether anything outside reaches it, the signal.

Step 3 - reading what it printed

| Signal | What it means | What to do | |---|---|---| | active | Being worked on | Nothing | | stable | Quiet, and something still reaches it | Leave it alone - finished and in use | | unused - N months | Quiet, and nothing reaches it | The only real retire candidate |

When the AGE line says young, write one line directly above the table, with the day count and date from the AGE line: "This project is N days old. Nothing here can be quiet for six months yet, so the Signal column starts to mean something on DATE." Every active is then a fact about the calendar, not the code. On an ok line, write no such note.

Step 4 - the backlog, sourced and never invented

Three sources, each item naming its own: session reports in ~/.claude/open-steps/reports/<project>/ (⏳ deferred rows, "Anything needed from you", recorded debt); Step 3 (each unused part, one retire candidate); the user, by hand. Nothing else - a gap you noticed while reading the code is not a task, so say it in the chat and let the user decide.

Step 5 - the tracker, if there is one

An offer, never a quiet action, and only for the "What is next" rows.

  1. Find it, never install it. gh repo view --json hasIssuesEnabled -q .hasIssuesEnabled, then any tracker already connected. Nothing connected is a normal answer: say so under "What is next", then Step 6.
  2. Search before you propose: gh issue list --search '<a few words>' --state all --json number,title,state. A closed ticket is still a match; where the search did not run, say so.
  3. Ask, then create. Plain lines through os-ask-simple; on a yes create those and nothing else, and only where this pass measured what the item rests on. Write each number back into its row, same pass.

Step 6 - plain words, before it is written

Put every sentence in the block through the os-say-simple rules first: one sentence one idea, active voice, one word for one thing. That covers the prose, the "What it does" column and the backlog items - never a measured cell.

The shape

<!-- open-steps:begin -->
_Measured <date>. Stages are dated where they stand; anything undated in this
file is not measured._

## What this is

<Three or four sentences. What the product does, who runs it, what it runs
on. No jargon - a reader who has never seen the code.>

## What is in it

| Feature | What it does | Stage | Last worked on | Signal |
|---|---|---|---|---|
| <plain name> | <one line> | live (1 Sep) | 1 Sep | active |
| <plain name> | <one line> | not checked | 28 May | stable |

## Worth retiring

<One line each: what it is, how long unused, what would break. Empty is a
result - write "nothing found" and keep the heading.>

## What is next

- <item> - <source>
- <item> - <source> - #<ticket number, once one exists>

_Age and wiring measured <date>, fresh this pass. Stage comes from session
reports and is only as current as the date beside it._
<!-- open-steps:end -->

Stage is one of building, built - not shipped, live, retired or not checked, and it carries the date it came from, always - live (12 Jul), never bare live, and not checked never dated. When the newest date in that column is months behind the newest commit, the reports have stopped: say so.

Hard rules - these rules are the skill

  1. Measured and assumed never mix. Last worked on and Signal are script output every time. Stage has no measurement.
  2. Enumerate, never guess. The parts come from the script.
  3. Quiet is not dead. No retire candidate on age alone: quiet plus wired in is stable, which is what stops the map recommending you delete a working product.
  4. Never invent work. Every backlog item names its source.
  5. Never clobber. Only between the markers. Their file stays theirs.
  6. Recommend, never act. No deleting code, moving files, opening a pull request. "Worth retiring" is a sentence, not an action.
  7. Plain words. A feature is named as the user says it, not as the folder.
  8. One screen per section. Twelve features is a wall - group and say so.
  9. A ticket only after an explicit yes. Rule 6 is the code, this is everyone else's inbox.
  10. The plain-words pass never touches a measured cell.
  11. Re-measure, never quote. A number read out of the file is a number about the past wearing today's date.
  12. Every claim carries its date or says it has none. A map that cannot go visibly stale will go invisibly stale.

Why the rules are there, the traps the script cannot catch, and a worked example: references/.

Related Skills

View on GitHub
GitHub Stars902
CategoryProduct
Updated2mo ago
Forks87

Languages

Shell

Trust signals

100/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

No cautions
os-big-picture — Universal Skill: Install & Safety Check | SkillAgent