**Motivation:**

Minor updates to the upgrade script for testnet deploy

**Modifications:**

Update condition on whether to *not* continue upgrade. Previously was
hitting a zeus env var not found error.
 
**Result:**

Working upgrade scripts

…1687)

**Motivation:**

Address L-01 finding from the Certora audit: "Instant slasher setting
leaves stale slasher field in storage". When `_updateSlasher()` is
called with `instantEffectBlock=true`, only `pendingSlasher` and
`effectBlock` are updated, leaving the `slasher` field stale (address(0)
for new operator sets). While `getSlasher()` returns correct values, raw
storage is inconsistent.

**Modifications:**

- Updated `_updateSlasher()` in `AllocationManager.sol` to set
`params.slasher = slasher` when `instantEffectBlock=true`
- Updated NatSpec for `SlasherParams` struct in `IAllocationManager.sol`
to document storage behavior
- Added `getSlasherParams()` helper to `AllocationManagerHarness.sol`
for testing raw storage
- Added unit tests for instant slasher storage consistency
- Updated documentation in `docs/core/AllocationManager.md`

**Result:**

- Storage is now consistent when creating operator sets or migrating
slashers
- Both `slasher` and `pendingSlasher` fields are set immediately when
`instantEffectBlock=true`
- Raw storage reads will return expected values matching `getSlasher()`
return value

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

…operators (#1688)

**Motivation:**

Addresses audit finding **L-01: State inconsistency for new operator
allocation delay**. When a new operator registers through
DelegationManager, their allocation delay is set via
`_setAllocationDelay(..., newlyRegistered=true)`. Previously, only
`pendingDelay` and `effectBlock` were updated, leaving `delay` and
`isSet` at default values until applied. This created temporary storage
inconsistency.

**Modifications:**

- `AllocationManager._setAllocationDelay()`: When
`newlyRegistered=true`, now also sets `info.delay = delay` and
`info.isSet = true` for immediate storage consistency
- `IAllocationManager`: Updated NatSpec for `AllocationDelayInfo` struct
to document the immediate update behavior
- `AllocationManagerHarness`: Added `getAllocationDelayInfoRaw()`
function for testing raw storage values
- Unit tests: Added
`test_setAllocationDelay_newlyRegistered_setsDelayAndIsSetImmediately`
and `test_setAllocationDelay_existingOperator_delayNotSetImmediately`
- Documentation: Updated `setAllocationDelay` Effects section in
`docs/core/AllocationManager.md`

**Result:**

Raw storage in `_allocationDelayInfo[operator]` now correctly reflects
the effective allocation delay immediately for newly registered
operators. The `getAllocationDelay()` view function behavior remains
unchanged (it already returned correct values by applying pending values
in-memory).

🤖 Generated with [Claude Code](https://claude.com/claude-code)

---------

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

…1689)

**Motivation:**

Address informational findings from the Certora audit of EigenLayer
Slashing UX Improvements (PR-1645/PR-544). These are quality-of-life and
documentation improvements that enhance code clarity and prevent
potential edge case issues.

**Modifications:**

1. **Fix I-01: Re-proposing same pending slasher is now a no-op**
- Added check in `_updateSlasher()` to skip processing if the proposed
slasher is already pending and hasn't taken effect
- Prevents accidentally restarting the delay countdown when re-proposing
the same slasher

2. **Fix I-02: Add NatSpec documentation for
getSlasher/getPendingSlasher**
- Updated interface and implementation NatSpec to document that these
functions return `address(0)`/`0` for non-existent operator sets
   - Helps callers understand expected behavior without validation

3. **Fix I-03: Add separate SLASHER_CONFIGURATION_DELAY constant**
- Added new `SLASHER_CONFIGURATION_DELAY` immutable in
`AllocationManagerStorage`
- Currently set to same value as `ALLOCATION_CONFIGURATION_DELAY` but
can be changed independently in future upgrades
   - Updated `_updateSlasher()` to use the new constant
   - Added interface getter

4. **Fix I-05: Add gas warning documentation for migrateSlashers**
   - Added NatSpec warning about O(appointees) gas cost per operator set
- `PermissionController.getAppointees()` enumerates full appointee set
which can be expensive
- Documented that large appointee sets may cause block gas limit issues

**Result:**

- `updateSlasher` is idempotent when re-proposing the same slasher
- Clearer documentation on view function return values
- Slasher delay can be configured independently in future upgrades
- Users are warned about potential gas issues with `migrateSlashers`

🤖 Generated with [Claude Code](https://claude.com/claude-code)

---------

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

**Motivation:**

Address informational findings from the Certora audit for Protocol
Registry (PR-1655). These improvements enhance validation, prevent
potential issues with orphaned configurations, and improve documentation
clarity.

**Modifications:**

1. **Fix I-01: ship() lacks validation**
- Added array length validation: `addresses.length == configs.length ==
names.length`
- Added zero address validation: revert if any address is `address(0)`
   - Added new errors: `ArrayLengthMismatch()` and `InputAddressZero()`

2. **Fix I-02: Orphaned configs on name overwrite**
- When re-shipping a name with a new address, the old address's
`DeploymentConfig` is now deleted
   - Added `DeploymentConfigDeleted(address)` event to signal cleanup
   - Prevents orphaned configs that could cause confusion

3. **Fix I-03: configure() for unshipped addresses**
   - Update `configure` function to pass in a name
- Added validation that the address must be a shipped deployment before
allowing configuration

5. **Fix I-04: Fix misleading NatSpec**
- Updated `ship()` NatSpec to document that names can be re-shipped with
new addresses
- Clarified that old address configs are automatically deleted when this
happens
- Updated `configure()` NatSpec to indicate address must be previously
shipped

6. **Fix I-05: Document pauseAll blocking**
- Added warning documentation that `pauseAll()` will revert if ANY
pausable deployment fails
- A single misconfigured deployment can block the entire protocol pause
   - Important for operational awareness during emergency scenarios

**Result:**

- `ship()` validates inputs before processing
- No more orphaned configs when re-shipping names
- `configure()` only works for shipped addresses
- Clearer documentation on expected behavior and edge cases
- Operators understand risks of `pauseAll()` failure scenarios

🤖 Generated with [Claude Code](https://claude.com/claude-code)

---------

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>