src/ui/key_handler/enter_handlers.rs (1)

1077-1080: Consider defaulting solver permission to Read for safer UX.

Line 1079 currently defaults to ReadWrite; starting at Read would better follow least-privilege and reduce accidental over-granting.

Proposed tweak

-                    permission: crate::ui::SolverPermission::ReadWrite,
+                    permission: crate::ui::SolverPermission::Read,

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/ui/key_handler/enter_handlers.rs` around lines 1077 - 1080, The code
currently sets the new AddSolver state with permission
SolverPermission::ReadWrite; change this default to the least-privilege option
by replacing SolverPermission::ReadWrite with SolverPermission::Read when
constructing UiMode::AdminMode(AdminMode::AddSolver(crate::ui::AddSolverState {
key_input: key_state, permission: ... })); ensure you only change the permission
enum value in the AddSolverState construction (leave UiMode::AdminMode,
AdminMode::AddSolver, and key_input as-is).

src/ui/help_popup.rs (1)

241-241: Include the selection keybinding in this help text.

Line 241 explains what to choose, but adding Left/Right would also explain how to choose it in-popup.

Suggested wording

-            "Enter solver npub and choose read or read-write permission before confirming.",
+            "Enter solver npub, use Left/Right to choose read or read-write, then confirm.",

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/ui/help_popup.rs` at line 241, Update the help text string in
src/ui/help_popup.rs (the line containing "Enter solver npub and choose read or
read-write permission before confirming.") to include the selection keybinding,
e.g., append "Use Left/Right to switch" or similar so users know how to toggle
the read vs read-write option; modify the exact literal in the help popup
rendering code where this message is defined to include "Left/Right" (or
"Left/Right arrow keys") next to the instruction.

src/util/order_utils/execute_admin_add_solver.rs (2)

102-114: Add a hex-input regression test for payload generation.

Current tests validate npub cases only. Add one case with a 64-char hex key to verify normalization + :read suffix behavior, since that path is part of this new payload logic.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/util/order_utils/execute_admin_add_solver.rs` around lines 102 - 114, Add
a new test that covers the hex-key normalization path: call
build_solver_payload_text with a 64-character hex key string and
SolverPermission::Read and assert the returned payload equals "{hex}:read"
(i.e., the hex key plus ":read"); also add a complementary test that calls
build_solver_payload_text with the same hex key and SolverPermission::ReadWrite
and assert the payload equals the hex key unchanged. Place these tests alongside
the existing add_solver_payload_* tests that use npub so the hex-input
regression for build_solver_payload_text is exercised.

---

8-8: Consider decoupling order_utils from crate::ui types.

src/util/order_utils/execute_admin_add_solver.rs importing crate::ui::SolverPermission couples utility/protocol code to UI state. Prefer moving this enum to a shared/domain module consumed by both layers.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/util/order_utils/execute_admin_add_solver.rs` at line 8, The file
execute_admin_add_solver.rs currently imports SolverPermission from crate::ui
which couples utility code to the UI; move the SolverPermission enum into a
shared/domain module (e.g., crate::domain::permissions or crate::shared::types)
and update execute_admin_add_solver.rs to import the enum from that new module
instead of crate::ui; update any other files that reference
crate::ui::SolverPermission to use the new path and run tests/compile to ensure
no remaining crate::ui references remain (look for the symbol SolverPermission
and the execute_admin_add_solver.rs import line).

src/ui/key_handler/admin_handlers.rs (1)

92-110: Optional: remove extra solver_pubkey clone before spawn.

You can move solver_pubkey into the async block directly and borrow it there, avoiding the intermediate clone.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/ui/key_handler/admin_handlers.rs` around lines 92 - 110, There’s an
unnecessary clone of solver_pubkey (solver_pubkey_clone) before spawning the
task; instead move/own solver_pubkey into the async block and use it there to
avoid the extra clone. Replace usage of solver_pubkey_clone with solver_pubkey
inside the tokio::spawn closure and ensure execute_admin_add_solver is called
with a reference or moved value consistent with its signature; remove the
solver_pubkey_clone binding and adjust the surrounding code that prepares
parameters (client_clone, result_tx, mostro_info) so they are moved/borrowed
into the spawned task similarly.

src/main.rs (1)

103-123: Optional: extract shared paste-append helper for admin key inputs.

Line 105-123 repeats the same sanitize/append/just_pasted pattern for two admin input states; pulling this into a small helper would reduce drift risk on future input updates.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/main.rs` around lines 103 - 123, Extract the repeated paste-and-append
logic into a small helper function (e.g., sanitize_and_append_paste) and replace
the duplicated blocks in the UiMode::AdminMode(AdminMode::AddSolver(ref mut
add_solver_state)) and UiMode::AdminMode(AdminMode::SetupAdminKey(ref mut
key_state)) arms to call it; the helper should take the pasted_text and a
mutable reference to the input buffer and just_pasted flag (or the state
struct), perform the chars().filter(|c| !c.is_control() || *c ==
'\t').collect(), append the result to the target key_input (for AddSolver use
add_solver_state.key_input.key_input.push_str, for SetupAdminKey use
key_state.key_input.push_str), and set just_pasted = true so both
add_solver_state and key_state reuse the same logic.

src/ui/draw.rs (1)

460-464: Optional render-path optimization: avoid cloning solver input each draw.

You can render with &str/Cow<'_, str> instead of allocating a new String every frame on Line 463.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/ui/draw.rs` around lines 460 - 464, The current code always allocates a
new String for input_display by cloning add_solver_state.key_input.key_input;
instead use a borrowed &str or Cow<'_, str> to avoid per-frame allocation:
change input_display's type to &str or Cow<'_, str> and return either a borrowed
literal ("npub... / hex...") or a borrowed slice of
add_solver_state.key_input.key_input (e.g.,
&add_solver_state.key_input.key_input) instead of calling .clone(), referencing
the symbol add_solver_state.key_input.key_input and the binding input_display to
locate where to update.