[thebrandonlucas]

Fixes DEV-2903

Problem

#22 made a missing --org, --env, or --wallet open a picker, but only when a command could not run without it. You still could not browse: voltage wallets get without an ID failed with wallets get needs a wallet ID … voltage wallets list shows the IDs, and there was no way to look at one organization or environment at all. You still had to copy UUIDs out of a list by hand.

Behavior

On a terminal, a read that names one resource but has no ID now picks it with the existing picker, then shows it:

Command, on a terminal	Before	After
wallets get, wallets ledger list, wallets policies get, wallets payments summary without WALLET_ID/--wallet	exit 2, needs a wallet ID	asks for the org if missing, then Wallet:, then shows that wallet
organizations get (new)	did not exist	shows the org that --org or the profile selects, or asks Organization:
environments get (new)	did not exist	shows the env that --env or the profile selects, or asks Organization:/Environment:
wallets update/delete without an ID	exit 2	unchanged: exit 2 before any prompt or request

After a pick, the CLI prints the usual Next time, pass --wallet ID … notice. Esc ends with the same error as before.

Nothing changes for scripts (DEV-2871). Off a terminal, with --json, --output json|ndjson, or --no-input, nothing is asked. A wallet read without an ID still fails with exit 2 and the same message and hint, before any credential is resolved or request is sent. organizations get and environments get without a selection report the usual --org is required and Exactly one --env is required errors.

Why this design

• The gap was in get. Every command that needs scope already picks it, and the one command that failed for a missing ID with a "go copy it from the list" hint was a wallet read. So I extended that rather than making list interactive or adding a copy-to-clipboard step. A list stays a list, which is safe to pipe.
• Reads only. A change (update, delete) still has to name its wallet. That keeps the existing rule that a delete is never aimed by ambient state, and keeps a picker out of the path of a change.
• organizations get and environments get take no positional ID. They show the organization or environment the scope selects, so --org, --env, and profiles work as they do in every other command. The pickers then come for free from pick::needs. The auth service has no single-resource route, so get finds the item in the same listing that list uses, and reports Organization ID is not visible to this login when it is not there.
• A wallet read never takes its wallet from VOLTAGE_WALLET_ID. Same rule as feat(cli): pick a missing organization, environment, or wallet #22's --wallet change. The picked wallet goes into the command's own ID, not the scope (pick::Picked.wallet).

Code

• cli.rs: a wallet operation without an ID now parses to resource_id: None instead of failing in clap validation. The error moves to cli::missing_wallet, which pick raises. ListCommand becomes DiscoveryCommand { List, Get }, with help examples for the new commands and for wallets get.
• pick.rs: wallet_target decides between picking and stopping (stopping for a change, off a terminal, or for JSON). ask now returns Picked { flags, wallet }, so fill can put the wallet in the scope or in the command's own ID.
• startup.rs: discover serves list and get. auth::Resolver::discover names the subcommand in its API-key error (organizations get needs a browser login …).
• Docs: guides/login-and-profiles.md (pickers browse), guides/wallets.md, generated commands.md, and a CHANGELOG line.

Checks

• Branch updated with a merge of master (CHANGELOG conflict: kept both lines). On the merged head: cargo fmt --check, cargo clippy --locked --all-targets -- -D warnings, cargo machete --with-metadata, cargo test --locked (200 unit, 152 integration), python3 scripts/check-coverage.py: all pass.
• New tests:
  • On a PTY: pick a wallet and show it; Esc; --json and --no-input (no prompt, no request); wallets delete (no prompt); pick an org for organizations get; pick an env for environments get; --json with no env (no prompt, no request).
  • Off a terminal: wallets ledger list with no ID; organizations get with a visible org, an org that is not visible, and no org; the API-key error for organizations get.
  • Unit: wallet_target for each read, change, and terminal/script case, and the new needs.
• CRAP (scripts/crap-report.sh, on the merged head): 0 of 851 measured functions at or above 20. Highest overall 18.10 (api.rs wait_for_payment, unchanged). Highest changed: pick::ask 14.00 and pick::fill 13.00, both 100% covered; auth::Resolver::discover 8, startup::discover 7, pick::wallet_target 5. Line coverage 97.97%.
• Diff reviewed against origin/master with voltage-pr-review and the Rust style checklist.
• Not run: staging. The steps below are for the reviewer.

Try it on staging

alias voltage="$PWD/scripts/voltage-staging"
voltage --build
voltage login --account me --credential-store file

voltage organizations get --account me

Expect: an Organization: picker. Type to filter, press Enter. The organization's fields follow, and then Next time, pass --org ORG_ID …. With only one organization, it is taken without asking.

voltage environments get --account me --org ORG_ID
voltage wallets get --account me --org ORG_ID

Expect: an Environment: picker, then the environment. Then a Wallet: picker (name CURRENCY UUID), then the wallet card with balance and activity, as for wallets get WALLET_ID.

voltage wallets get --account me --org ORG_ID --json; echo $?
voltage wallets get --account me --org ORG_ID | cat; echo $?
voltage wallets delete --account me --org ORG_ID; echo $?

Expect: no picker in any of these. wallets get needs a wallet ID (or wallets delete needs a wallet ID), with the hint, and exit code 2.

Merge order with #127

#127 now sits directly below this PR in stack #135, so the picker's pseudo-terminal tests (pick, Esc) run against #127's new terminal guard; the whole suite passes on this head. Merging the layers below also gave DiscoveryCommand::List #120's --all (and updated #120's parse test to DiscoveryCommand::List) and kept #129's context --format env dispatch beside this PR's discover.

Open questions

• Should changes (wallets update/delete) pick too? They are describe-only without --execute, but I kept them explicit to stay conservative.
• Other single-resource reads (payments get, webhooks get, quotes get) could pick the same way. I left them out: their lists can be long and paginated, and they have no names to filter on.

Stack

Stack #135, layer 6 of 9 (bottom targets master):

#	PR	Change
1	#97	Staging setup and a Try it section for each PR
2	#120	--all on organizations, environments, and profiles list
3	#128	Warn when a profile or flag overrides a set VOLTAGE_* variable
4	#129	voltage context --format env
5	#127	Replace hand-rolled terminal code with crossterm
6	#126 (this PR)	Pick the wallet, organization, or environment to show
7	#130	Prefilled GitHub issue links from voltage feedback and panics
8	#123	voltage environments create
9	#131	Replace flake.nix with a Kaifile

The layers are chained with merge commits, because history is not rewritten. Use squash and merge, bottom first.

[tee8z]

Approved at 51e7afb. Moving the missing-wallet error from parsing to pick::fill is safe: Session::open calls fill on every path, and wallet_target raises missing_wallet off a terminal, for JSON, and for every non-GET wallet operation, before any credential is resolved or request is sent. The picked wallet goes to resource_id, never the scope, so VOLTAGE_WALLET_ID still cannot aim a wallet command. organizations get and environments get reuse pick::needs, and the not-visible path is tested.

Non-blocking: discover finds the item by comparing id.to_string() with the API string. That depends on the API returning lowercase UUIDs. Parsing item["id"] as a Uuid would not.

CRAP: not run on this head yet; I will add the result when it finishes. CI is green.

[tee8z]

CRAP on 51e7afb: 875 of 1093 functions measured, 0 violations. Highest: 18.10 wait_for_payment (unchanged), then 18.00 endpoint (unchanged).

[thebrandonlucas]

Took the non-blocking suggestion in 8d9618b. discover now parses item["id"] as a Uuid before it compares. organizations_get_shows_the_selected_organization_and_never_asks_a_script now lists an ID in capitals and gets it by its lowercase form. That test fails without the change. CRAP on 8d9618b: 0 violations. The push may have dismissed your approval.

[tee8z]

Reviewed 49590cc against #127 (5f64b27). No blocking findings in this layer.

Wallet reads can pick their explicit target; wallet mutations still fail before discovery when the target is absent. The PTY tests exercise wallet/org/env selection, Esc, and the machine-output/no-input paths.

Local gate passed: formatting, Clippy with warnings denied, dependency hygiene, full tests (210 unit + 156 integration), and contract/docs generation checks. Required CRAP check passed: 877/1097 functions measured, 220 unmeasured, no scores >=20; highest 18.10 (wait_for_payment).

Limits: Linux mock-service/PTY validation; no live staging requests or manual Windows/macOS run. Approval applies to this incremental stack layer.