security: [T1-07] reject oracle_id > 255 to prevent truncation (DGB-SEC-004)#376
Merged
JaredTate merged 8 commits intofeature/digidollar-v1from Feb 13, 2026
Conversation
…EC-004) The on-chain script format stores oracle_id as a single byte (uint8). The in-memory type is uint32_t. Without validation, IDs > 255 are silently truncated via `& 0xFF`, causing ID collisions (e.g. oracle 1 and 257 become indistinguishable) and signature verification failures. Fix adds validation at two layers: - IsValidOracleMessage(): rejects at P2P ingestion boundary - CreateOracleScript(): defense-in-depth rejection at serialization for both Phase One and Phase Two script formats
There was a problem hiding this comment.
Pull request overview
This PR addresses a critical security vulnerability (DGB-SEC-004) where oracle IDs are stored as uint32_t in memory but serialized as a single byte in on-chain scripts. IDs greater than 255 would be silently truncated via & 0xFF, causing ID collisions and signature verification failures. The fix implements defense-in-depth validation at three layers to reject oracle_id > 255.
Changes:
- Added validation in
IsValidOracleMessage()to reject oracle messages with ID > 255 at P2P ingestion - Added defense-in-depth checks in
CreateOracleScript()for both Phase One and Phase Two serialization paths - All checks include clear security comments referencing DGB-SEC-004
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
OpenSSL 1.1.1w Configure misparses multi-word flags like "-arch arm64" when passed as positional args — the shell splits them and Configure treats the second word as a target name, causing "target already defined" on ARM64 macOS. Pass CFLAGS/CPPFLAGS as VAR=value assignments per OpenSSL's official INSTALL docs. libcurl 8.5.0 fopen.c uses fileno/fdopen (POSIX) which are hidden by the depends system's strict -std=c11. Add -D_GNU_SOURCE for Linux.
8 tasks
Use $$ deferred evaluation for CFLAGS/CPPFLAGS assignments so OS-specific flags (cflags_linux, cppflags_linux) appended by funcs.mk after set_vars are included. Route -fPIC and -D_GNU_SOURCE through proper cflags/cppflags variables instead of bare config_opts to avoid OpenSSL's "Mixing make variables" rejection.
The depends system never defines $(package)_arflags, so ARFLAGS=$($(package)_arflags) in config_env exported an empty ARFLAGS="" to the environment. When CFLAGS/CPPFLAGS are passed as VAR=value assignments (not positional flags), OpenSSL 1.1.1w's Configure sets $anyuseradd=false and falls back to reading env vars. The empty ARFLAGS overrides the target default "r" from Configurations/00-base-templates.conf, producing a Makefile with ARFLAGS= (empty). This causes "ar: two different operation options specified" on Linux and "ar: illegal option -- /" on macOS during build_libs, as ar interprets the archive path as flags. Fix: remove ARFLAGS from config_env entirely, letting Configure use its target default ARFLAGS="r". Verified locally: - Linux x86_64: ar r apps/libapps.a ... (ARFLAGS=r in Makefile) - macOS ARM64: ar r apps/libapps.a ... (ARFLAGS=r in Makefile) - Full build_libs completes successfully on darwin64-arm64-cc
Replace BOOST_CHECK with BOOST_CHECK_MESSAGE in 5 mint validation tests that fail on CI but pass locally. Diagnostic output captures the exact reject reason, script sizes, collateral amounts, and script type identification to help debug the CI-specific failure.
Replace ClearFreeze() with ClearHistory() in the validation test fixture. ClearFreeze() only clears freeze flags, but UpdateState() (called during ValidateDigiDollarTransaction) recalculates volatility from stale price history left by earlier test suites (e.g., digidollar_health_tests) and re-sets the freeze. This caused all "valid mint" tests to fail with "minting-frozen-volatility" on CI where the test suite ordering (linker-dependent) places health_tests before validation_tests. ClearHistory() resets price history, volatility state, and all freeze flags, ensuring each test starts from a clean state.
The merge with feature/digidollar-v1 silently dropped critical code: 1. DD amount double-counting prevention in ValidateMintTransaction() The OP_RETURN and DD token output both encode ddAmount. Without the consistency check, totalDD was counted twice (e.g. 10000 + 10000 = 20000), requiring 2x the collateral and failing all valid mint tests. 2. Missing #include <digidollar/health.h> 3. Improved log format for insufficient-collateral diagnostic Also adopts the base branch's test file which includes: - DD OP_RETURN outputs required for T1-04b NUMS verification - Proper NUMS key usage via GetCollateralNUMSKey() - ClearHistory() fix for cross-suite volatility state pollution
|
Reviewed and merged manually into What was applied:All three defense-in-depth checks from
This prevents the silent What was NOT applied:
Verification:
Closing as manually merged. Good catch @gto90 — this was a real gap in our input validation. |
4 tasks
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
3 participants
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.
Summary
uint32_tin-memory but serialized as 1 byte in on-chain scripts via& 0xFF. IDs > 255 are silently truncated, causing ID collisions and signature verification failuresIsValidOracleMessage) and script serialization (CreateOracleScript, both Phase One and Phase Two)Test plan
IsValidOracleMessage()test_digibyte --run_test=oracle_tests