chore(preflight): bump skill integrity pin for v2 (NEAR-Intents cross-check) - #247
Merged
Merged
Conversation
…-check)
The vaultpilot-preflight skill (separate repo, lives under
~/.claude/skills/) gains an agent-side cross-check for LiFi
intermediate-chain bridges in v2: the agent now consults its own
hardcoded copy of the (bridge name, intermediate chain ID) allowlist
when verifying cross-chain bridge calldata, so a tampered MCP that
secretly widens its own allowlist gets caught.
The skill content change shifts both the SHA-256 of SKILL.md and the
in-file integrity sentinel. Update the pin in this MCP's source so the
on-each-flow integrity check the MCP delegates to the agent (see same
file, ~30 lines above) succeeds against the new skill version.
- SHA-256:
10db66ad5fcfa250b4a5a753fd745eac4dc358b41ccfc1a1258f67040b7eddea (old, v1)
→ fc4a40469e11156e7cccdc1539b93e430b60a49bf75d01b41de112d8c301f1a8 (new, v2)
- Sentinel fragment B: _v1_ → _v2_
- Sentinel fragment C: 7780bfeee9a49f01 → 43b1d2403a0c2a94
Coordinated with the matching PR in szhygulin/vaultpilot-skill.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Coordinated with `vaultpilot-skill#3`.
Summary
The `vaultpilot-preflight` skill picks up an agent-side cross-check for LiFi intermediate-chain bridges (Invariant #6 in the skill PR — agent independently verifies `BridgeData.destinationChainId` against a hardcoded list of known intermediate-chain bridges, with NEAR Intents at chain ID `1885080386571452` as the only entry today). That content change shifts both:
This MCP pins both values in source (`src/index.ts` ~line 1161 and ~line 1167). The pin must be bumped to the new values, otherwise every `prepare_` / `preview_` flow will fail the integrity check until the next release of this MCP.
Why two copies of the allowlist
PR #239 added `INTERMEDIATE_CHAIN_BRIDGES` to this MCP (`src/modules/swap/intermediate-chain-bridges.ts`). That alone has the well-known limitation: a compromised MCP can rewrite the constant. The skill carries an independent copy, lives outside this MCP's reach (separate repo, separate npm package, lives under `~/.claude/skills/`), and the agent is now instructed to cross-check both. Single-side tamper attempts get caught.
Test plan
🤖 Generated with Claude Code