You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Convert versioned checker reports to private SARIF and NDJSON - #1
Add an optional standalone offline converter for canonical v1.0 reports and six explicitly versioned legacy checker profiles. Produce SARIF 2.1.0 or NDJSON with stable observation fingerprints, conflict-preserving deduplication, counts and completion state. Private strings are projected to hashes by default; relative source locations require opt-in. Arbitrary tool version labels are rejected, and canonical limitations/snapshots contribute to evidence identity.
Validation: 13 local stdlib tests pass; five SARIF variants pass the vendored official OASIS schema. Actual synthetic outputs from the supported checker CLIs were accepted. Copilot hosted findings were addressed, followed by an independent GitHub Copilot CLI review of 0c8e2c5 with no actionable bridge finding. No targets, uploads or runtime assessment. This supports node2shell-check issue #14; each checker remains independently usable.
Adds an optional, standalone offline bridge for the checker family, converting existing versioned reports into privacy-projected SARIF or NDJSON without adding checker dependencies or uploads.
Changes:
Bundles canonical validation and six explicit legacy adapters.
The reason will be displayed to describe this comment to others. Learn more.
🔵 Needs a closer look
Unresolved privacy and conversion correctness issues require correction and human confirmation, with exact-head hosted CI still unverified.
No open issues 🎉
3 previously missed issues
In code that hasn't changed since last review
Protect free-form tool versions and completion metadata from disclosure
scan_results_bridge.py:654
The canonical contract allows arbitrary text in tool.version, so a value such as build-token=SYNTHETIC_SECRET bypasses privacy projection and appears verbatim in both default output formats. Hash free-form tool versions or restrict them to an explicitly public format before exporting them. Apply the same handling to the completion keys at lines 765–766, which also populate SARIF tool versions for empty reports.
Deduplicate evidence locations before determining unique source position
scan_results_bridge.py:691
Distinct evidence IDs can legitimately refer to the same source path and line. With two such entries, this list has length 2, so the converter drops the unambiguous location and changes the fingerprint even though the asset, component and source position are unchanged. Deduplicate path/line pairs before checking for a single location, and add a regression case with two evidence entries at the same position.
Set explicit SARIF evaluation kind for non-finding observations
scan_results_bridge.py:726
Omitting kind makes every result inherit SARIF's default fail, including unknown, fixed_version_reported, and mitigated observations. Setting level to note only changes severity; SARIF consumers still see these as failed rule evaluations unless they understand the custom status property. Set an explicit evaluation kind for non-finding observations, such as review for uncertainty or informational with a compatible level for metadata-only results.
The reason will be displayed to describe this comment to others. Learn more.
🔵 Needs a closer look
Unresolved privacy and observation-identity issues can expose private metadata or lose meaningful report distinctions.
No open issues 🎉
3 previously missed issues
In code that hasn't changed since last review
Prevent Private Tool Versions from Leaking into Published Metadata
scan_results_bridge.py:665
The canonical contract allows arbitrary text in tool.version, but this value bypasses the privacy projection. A valid report whose version is https://private.invalid/?token=SYNTHETIC_SECRET exits successfully and publishes that URL/token unchanged in SARIF driver metadata and NDJSON result rows; even an empty SARIF report exposes it through the completion keys. Validate published versions against an explicitly public format or privacy-project free-form versions consistently in both normalized rows and run/completion metadata before serialization.
Deduplicate Evidence Locations Before Determining Location Ambiguity
scan_results_bridge.py:702
Two valid source-evidence entries with different evidence IDs can refer to the same path and line. Counting entries rather than distinct locations treats this unambiguous case as locationless: SARIF omits the location even with path publication enabled, and adding corroborating evidence changes the fingerprint despite the source location remaining unchanged. Accept the location when all valid entries refer to the same path/line.
Include Report Limitations in Canonical Evidence Digest
scan_results_bridge.py:717
Report-level limitations are omitted from canonical evidence and its digest. Consequently, complete reports with identical results and snapshots but different coverage limitations normalize identically and are merged as exact duplicates, losing the distinction between their assessment constraints. Include the limitations before privacy projection and hashing so they remain privately represented and observations with different limitations stay distinct.
Final gate: exact head 0c8e2c5 has SUCCESS hosted CI across Linux, Windows and macOS. GitHub Copilot CLI independently reviewed its complete source/test/doc/workflow diff (excluding the vendored SARIF schema), with no actionable bridge finding. The provenance URL was independently checked: CompleTetech-Security is the actual organization. All 13 local tests, five official-schema SARIF probes and diff/checksum validation pass. Privacy hashes are linkable/guessable and numbers remain as documented; no anonymity or production assessment is claimed.
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
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.
Add an optional standalone offline converter for canonical v1.0 reports and six explicitly versioned legacy checker profiles. Produce SARIF 2.1.0 or NDJSON with stable observation fingerprints, conflict-preserving deduplication, counts and completion state. Private strings are projected to hashes by default; relative source locations require opt-in. Arbitrary tool version labels are rejected, and canonical limitations/snapshots contribute to evidence identity.
Validation: 13 local stdlib tests pass; five SARIF variants pass the vendored official OASIS schema. Actual synthetic outputs from the supported checker CLIs were accepted. Copilot hosted findings were addressed, followed by an independent GitHub Copilot CLI review of 0c8e2c5 with no actionable bridge finding. No targets, uploads or runtime assessment. This supports node2shell-check issue #14; each checker remains independently usable.