A skill's text calls its command by its full name

.agents/skills/browser/SKILL.md

@@ -5,24 +5,24 @@ description: A real browser you drive from the shell, to see or act on a web pag
 
 # The browser
 
-A headless Chrome of your own, on an empty profile: nobody's logins are in it. Drive it with the `browser` command, a dependency of this repository (`@openagt/skill-browser`), run as `npx browser`. When that fails for a missing `node_modules`, install with the lockfile's package manager (`npm install` for `package-lock.json`) and run it again. It needs Chrome on the machine, or `CHROME_PATH` set to a Chrome or Chromium executable.
+A headless Chrome of your own, on an empty profile: nobody's logins are in it. Drive it with the `browser` command, always run by its full name: `npx @openagt/skill-browser@0.1`. It needs Chrome on the machine, or `CHROME_PATH` set to a Chrome or Chromium executable.
 
 ```
-npx browser open <address>       open the address, starting the browser if none is open
-npx browser read                 print the page again
-npx browser click <n>            click element n
-npx browser type <n> <text>      replace what element n holds with the text; on a select, pick that option;
-                                 on a date or time field, give the value as 2024-01-31, 13:45, 2024-01
-npx browser press <key>          Enter, Tab, Escape, Backspace, Space, ArrowUp, ArrowDown, ArrowLeft, ArrowRight
-npx browser screenshot [file]    save what the page shows as a PNG and print its path (a temporary file
-                                 when none is named; never name one inside the repository)
-npx browser eval <script>        run a JavaScript expression in the page and print its value, as JSON
-                                 (`document.title`, not `return document.title`; `await` works; an
-                                 element prints as `{}`, so ask for its properties)
-npx browser close                close the browser
+npx @openagt/skill-browser@0.1 open <address>       open the address, starting the browser if none is open
+npx @openagt/skill-browser@0.1 read                 print the page again
+npx @openagt/skill-browser@0.1 click <n>            click element n
+npx @openagt/skill-browser@0.1 type <n> <text>      replace what element n holds with the text; on a select, pick that option;
+                                                    on a date or time field, give the value as 2024-01-31, 13:45, 2024-01
+npx @openagt/skill-browser@0.1 press <key>          Enter, Tab, Escape, Backspace, Space, ArrowUp, ArrowDown, ArrowLeft, ArrowRight
+npx @openagt/skill-browser@0.1 screenshot [file]    save what the page shows as a PNG and print its path (a temporary file
+                                                    when none is named; never name one inside the repository)
+npx @openagt/skill-browser@0.1 eval <script>        run a JavaScript expression in the page and print its value, as JSON
+                                                    (`document.title`, not `return document.title`; `await` works; an
+                                                    element prints as `{}`, so ask for its properties)
+npx @openagt/skill-browser@0.1 close                close the browser
 ```
 
-Quote a `<text>` or `<script>` with spaces as one argument: `npx browser type 2 'Ada Lovelace'`.
+Quote a `<text>` or `<script>` with spaces as one argument: `npx @openagt/skill-browser@0.1 type 2 'Ada Lovelace'`.
 
 `open`, `read`, `click`, `type` and `press` print the page once it has loaded (or after 10 seconds, as it stands): its title, its address, its text (the first 10,000 characters), then its elements (links, buttons, fields; the first 300) numbered `[n]`. What is inside an iframe or a shadow root is not printed; reach it with `eval`. `click` and `type` take that number, from the last print: when the page changes, the numbers do too. The numbers are written on the page's elements as `data-browser-ref` attributes. An address without `http://` or `https://` gets `http://`; no other kind opens. A dialog the page opens (alert, confirm, prompt) is accepted at once, and the next print begins with a `Dialog, accepted:` line naming it.
 

.agents/skills/discord/SKILL.md

@@ -5,16 +5,16 @@ description: Post a message to the team's Discord channel from the shell, when a
 
 # Discord
 
-Post one message to a Discord channel with the `discord` command, a dependency of this repository (`@openagt/skill-discord`), run as `npx discord`. When that fails for a missing `node_modules`, install with the lockfile's package manager (`npm install` for `package-lock.json`) and run it again.
+Post one message to a Discord channel with the `discord` command, always run by its full name: `npx @openagt/skill-discord@0.1`.
 
 ```
-npx discord send "<message>"     post the message to this machine's webhook
-npx discord status               whether this machine has a webhook, and where it comes from
+npx @openagt/skill-discord@0.1 send "<message>"     post the message to this machine's webhook
+npx @openagt/skill-discord@0.1 status               whether this machine has a webhook, and where it comes from
 ```
 
 Quote the message as one argument. It is posted as a Discord message, so Discord's markdown formats it (`*`, `_`, backticks), up to 2,000 characters (a longer one is cut and says so); `@everyone` and other mentions in it do not ping anyone.
 
-The channel is the one whose webhook this machine was given, by a person: `npx discord setup <webhook>` saves it for this user on this machine, `npx discord setup --clear` forgets it, and `DISCORD_WEBHOOK` in the environment wins over the saved one. Never ask for the webhook, never print it, and never write it into the repository: anyone holding it can post to the channel.
+The channel is the one whose webhook this machine was given, by a person: `npx @openagt/skill-discord@0.1 setup <webhook>` saves it for this user on this machine, `npx @openagt/skill-discord@0.1 setup --clear` forgets it, and `DISCORD_WEBHOOK` in the environment wins over the saved one. Never ask for the webhook, never print it, and never write it into the repository: anyone holding it can post to the channel.
 
 `send`, `setup` and `status` print one JSON line on stdout. A refusal exits 1 with the reason on stderr: `no-webhook` (none on this machine: say so and go on, a person sets it), `not-posted` (Discord or the network refused it; the detail says which, never the URL), `empty`, and for `setup`, `invalid`. A wrong command line exits 2 with the usage on stderr.
 
@@ -23,5 +23,5 @@ The channel is the one whose webhook this machine was given, by a person: `npx d
 A tool that runs coding agents may run a line of the person's when a run ends, with the message in `$MESSAGE`. This is the line:
 
 ```
-npx discord send "$MESSAGE"
+npx @openagt/skill-discord@0.1 send "$MESSAGE"
 ```

.agents/skills/github/SKILL.md

@@ -9,14 +9,14 @@ The project's pull requests and issues are on GitHub. This skill is how you reac
 
 ## The command
 
-`github` is a dependency of this repository (`@openagt/skill-github`). If `node_modules` is missing, install with the lockfile's package manager (`npm install` for `package-lock.json`). Then run `npx github` inside your checkout. It prints JSON; a refusal exits 1 with a line on stderr; a wrong command line exits 2 with the usage.
+Run the `github` command inside your checkout, always by its full name: `npx @openagt/skill-github@0.1`. It prints JSON; a refusal exits 1 with a line on stderr; a wrong command line exits 2 with the usage.
 
 ## Your pull request
 
 Open a pull request only when your task or the person asks for one, or asks you to publish your work. Push your branch first (your branches skill says how), then:
 
 ```
-npx github open --title "<one line naming what the change does>" --body "<what changed, and why>"
+npx @openagt/skill-github@0.1 open --title "<one line naming what the change does>" --body "<what changed, and why>"
 ```
 
 It opens the pull request for the branch you are on and prints it in `request` (`number`, `url`). A branch that already has an open pull request gets no second one: that one is answered, with `existing` true. Add `--merge` when the work may land on its own: the pull request then merges once its checks pass, also where the repository does not allow auto-merge (`merge` in the answer says `auto-armed`, `merged` or `watching`). Add `--draft` instead for a pull request a person should look at first; with `--merge` too, `--draft` is dropped.
@@ -26,7 +26,7 @@ A pull request whose body has a line `Closes #<number>` closes that issue when i
 ## Reading pull requests and issues
 
 ```
-npx github requests [--branch <b>] [--state open|merged|all] [--since <iso>]
+npx @openagt/skill-github@0.1 requests [--branch <b>] [--state open|merged|all] [--since <iso>]
 ```
 
 The project's newest 50 pull requests, newest first, all states unless `--state`, only `<b>`'s with `--branch`; `--since` keeps those created since, or merged since with `--state merged`. Each has `number`, `url`, `state` (`open`, `merged`, `closed`), `title`, `draft`, `branch`, `head`, `createdAt`, and for a merged one `mergedAt` and `mergeCommit` (the commit it landed as).
@@ -36,7 +36,7 @@ For a pull request's discussion and diff, and for the project's issues, use `gh`
 ## Landing a pull request for a person
 
 ```
-npx github merge <number>
+npx @openagt/skill-github@0.1 merge <number>
 ```
 
 A draft is marked ready, then the merge is armed as `--merge` arms it: `outcome` says `auto-armed`, `merged` or `watching`; a request no longer open is refused as `not-open`. Only when whoever started you asked you to land it.

.agents/skills/logs/SKILL.md

@@ -7,21 +7,21 @@ description: The record of every run agents made on this project — what was as
 
 Every run an agent made on this project leaves a record on the branch `agent-data`, never on a code branch; your checkout does not contain it. A run is two files under `agents/<who>/`: the card, `<id>.json` — what was asked, the branch, the pull request, how it ended, what it cost — and the diary, `<id>.jsonl` — what the agent said along the way, its result.
 
-Read them with the `logs` command, a dependency of this repository (`@openagt/skill-logs`), run as `npx logs`. When that fails for a missing `node_modules`, install with the lockfile's package manager (`npm install` for `package-lock.json`) and run it again. You only read: the program that ran an agent records its run, at its end. The command's `delete` and `patch`, and its `--local` and `--full` flags, are for the dashboard that shows the runs; never use them. A refusal exits 1 with a line on stderr; a wrong command line exits 2 with the usage.
+Read them with the `logs` command, always run by its full name: `npx @openagt/skill-logs@0.1`. You only read: the program that ran an agent records its run, at its end. The command's `delete` and `patch`, and its `--local` and `--full` flags, are for the dashboard that shows the runs; never use them. A refusal exits 1 with a line on stderr; a wrong command line exits 2 with the usage.
 
 ## Read
 
 ```
-npx logs [--branch <name>] [--limit N]
-                                 the runs, newest first, as one JSON array of cards — the newest 20
-                                 unless --limit says otherwise; --branch keeps the runs on one branch
-npx logs show <id>               one run: its card, plus `diary`, the lines of what the agent
-                                 said, its result, how it ended, and what it cost
+npx @openagt/skill-logs@0.1 [--branch <name>] [--limit N]
+                                                    the runs, newest first, as one JSON array of cards — the newest 20
+                                                    unless --limit says otherwise; --branch keeps the runs on one branch
+npx @openagt/skill-logs@0.1 show <id>               one run: its card, plus `diary`, the lines of what the agent
+                                                    said, its result, how it ended, and what it cost
 ```
 
 ## Before you plan or work a ticket, read its earlier runs
 
-A ticket may have been worked before. Claiming it names who claimed it before you: a run's id, shaped like `2026-09-08T18-14-30-111Z`, read with `npx logs show <id>`, or else a branch, read with `npx logs --branch <name>`; no run on that branch means no record: read the branch itself. A `stopped` or `failed` run tells you what to avoid: read what its agent said before it ended. A `done` run with a `pr` means the work may already be there: read the pull request before doing it again.
+A ticket may have been worked before. Claiming it names who claimed it before you: a run's id, shaped like `2026-09-08T18-14-30-111Z`, read with `npx @openagt/skill-logs@0.1 show <id>`, or else a branch, read with `npx @openagt/skill-logs@0.1 --branch <name>`; no run on that branch means no record: read the branch itself. A `stopped` or `failed` run tells you what to avoid: read what its agent said before it ended. A `done` run with a `pr` means the work may already be there: read the pull request before doing it again.
 
 ## The card
 

.agents/skills/orchestration/SKILL.md

@@ -5,16 +5,16 @@ description: "Split your task across subagents: plan it, get the person's yes, s
 
 # Orchestration
 
-A subagent is another coding agent you start on one part of your task: it works alone, in its own checkout. Use the `orchestration` command, a dependency of this repository (`@openagt/skill-orchestration`), run as `npx orchestration`. When that fails for a missing `node_modules`, install with the lockfile's package manager (`npm install` for `package-lock.json`) and run it again.
+A subagent is another coding agent you start on one part of your task: it works alone, in its own checkout. Use the `orchestration` command, always run by its full name: `npx @openagt/skill-orchestration@0.1`.
 
 ```
-npx orchestration plan <file>    save the file as your plan; answers the `question` to ask the person
-npx orchestration start --level <simple|hard> "<task>"
-                                 start a subagent on the task; answers its id at once, without waiting for it
-npx orchestration list           your subagents, newest first, as one JSON array
-npx orchestration read <id>      one subagent: its `status`, and once it ended its `branch` and `result`, its last reply
-npx orchestration stop <id>      stop a subagent that is running
-npx orchestration land <id>      merge an ended subagent's branch into yours, then delete that branch
+npx @openagt/skill-orchestration@0.1 plan <file>    save the file as your plan; answers the `question` to ask the person
+npx @openagt/skill-orchestration@0.1 start --level <simple|hard> "<task>"
+                                                    start a subagent on the task; answers its id at once, without waiting for it
+npx @openagt/skill-orchestration@0.1 list           your subagents, newest first, as one JSON array
+npx @openagt/skill-orchestration@0.1 read <id>      one subagent: its `status`, and once it ended its `branch` and `result`, its last reply
+npx @openagt/skill-orchestration@0.1 stop <id>      stop a subagent that is running
+npx @openagt/skill-orchestration@0.1 land <id>      merge an ended subagent's branch into yours, then delete that branch
 ```
 
 A refusal exits 1 with a line on stderr saying why.

.agents/skills/plan-tickets/SKILL.md

@@ -6,7 +6,7 @@ schedule:
   every: 6h
   waits-for: when a ticket that is not waiting, claimed, in review or queued has no plan, or an outdated one
   when: |-
-    npx tickets list | jq --argjson q "$(npx queue)" '[.[] | select(((.planned | not) or .outdated) and .pr == null and (.waiting | not) and (.locked | not) and ((.file | rtrimstr(".md")) as $s | $q | any(contains($s)) | not)) | .file]'
+    npx @openagt/skill-tickets@0.1 list | jq --argjson q "$(npx @openagt/skill-queue@0.1)" '[.[] | select(((.planned | not) or .outdated) and .pr == null and (.waiting | not) and (.locked | not) and ((.file | rtrimstr(".md")) as $s | $q | any(contains($s)) | not)) | .file]'
 ---
 
 Queue plans for the tickets that need one. Nobody will answer you: never ask, decide yourself. You only queue work, you never do it: write no plan yourself, the only thing you change is the queue.

.agents/skills/post-merge-cleanup/SKILL.md

@@ -6,7 +6,7 @@ schedule:
   every: 1h
   waits-for: when a pull request merged since the last clean-up has no clean-up yet
   when: |-
-    gh pr list --state merged --limit 50 --json number,body --search "merged:>=$(npx logs --limit 200 | jq -r '[.[] | select(.intent == "/post-merge-cleanup" and .status == "done")][0].startedAt // (now - 86400 | todate)')" | jq '[.[] | select(.body | contains("Post-merge cleanup done.") | not) | .number]'
+    gh pr list --state merged --limit 50 --json number,body --search "merged:>=$(npx @openagt/skill-logs@0.1 --limit 200 | jq -r '[.[] | select(.intent == "/post-merge-cleanup" and .status == "done")][0].startedAt // (now - 86400 | todate)')" | jq '[.[] | select(.body | contains("Post-merge cleanup done.") | not) | .number]'
 ---
 
 Write what merged work decided and taught into the project's knowledge files. Nobody will answer you: never ask, decide yourself.

.agents/skills/queue/SKILL.md

@@ -7,22 +7,22 @@ description: Where the project's agent queue lives, how to read it and change it
 
 The agent queue (`TODO_AGENTS.md`) lives on the branch `agent-data`, never on a code branch; your checkout does not contain it. It lists every task agents will work on next, in the order they will be taken.
 
-Read and change it with the `queue` command, a dependency of this repository (`@openagt/skill-queue`), run as `npx queue`. When that fails for a missing `node_modules`, install with the lockfile's package manager (`npm install` for `package-lock.json`) and run it again. Every change it makes is one commit on the `agent-data` branch, pushed straight to the remote when the project shares its records. The command's `--local` and `--full` flags are for the dashboard that shows the queue; never use them. A refusal exits 1 with a line on stderr; a wrong command line exits 2 with the usage.
+Read and change it with the `queue` command, always run by its full name: `npx @openagt/skill-queue@0.1`. Every change it makes is one commit on the `agent-data` branch, pushed straight to the remote when the project shares its records. The command's `--local` and `--full` flags are for the dashboard that shows the queue; never use them. A refusal exits 1 with a line on stderr; a wrong command line exits 2 with the usage.
 
 ## Read
 
 ```
-npx queue                        the open entries, in order of work, as one JSON array
+npx @openagt/skill-queue@0.1                        the open entries, in order of work, as one JSON array
 ```
 
 ## Change
 
 ```
-npx queue add <text> [--priority N]
-                                 put an entry on the queue; --priority (0-10) places it at the end of
-                                 that section; without it, the entry goes at the end of the file
-npx queue done <entry>           remove an entry: one quoted argument, exactly as `npx queue`
-                                 printed it; done means deleted
+npx @openagt/skill-queue@0.1 add <text> [--priority N]
+                                                    put an entry on the queue; --priority (0-10) places it at the end of
+                                                    that section; without it, the entry goes at the end of the file
+npx @openagt/skill-queue@0.1 done <entry>           remove an entry: one quoted argument, exactly as `npx @openagt/skill-queue@0.1`
+                                                    printed it; done means deleted
 ```
 
 ## Format
@@ -44,4 +44,4 @@ npx queue done <entry>           remove an entry: one quoted argument, exactly a
 ...
 ```
 
-The queue lists *all* tasks AI will work on next, sorted by priority. Priority 10 is rarely used (e.g. critical production bugs) and is treated as the utmost priority. Within a priority, the first tasks have higher priority (they are the "next" tasks within that "priority queue"). A done entry is removed (`npx queue done`).
+The queue lists *all* tasks AI will work on next, sorted by priority. Priority 10 is rarely used (e.g. critical production bugs) and is treated as the utmost priority. Within a priority, the first tasks have higher priority (they are the "next" tasks within that "priority queue"). A done entry is removed (`npx @openagt/skill-queue@0.1 done`).