Add AVE (Agentic Vulnerability Enumeration) mappings, 4-record pilot - #66
Open
chaksaray wants to merge 1 commit into
Open
Add AVE (Agentic Vulnerability Enumeration) mappings, 4-record pilot#66chaksaray wants to merge 1 commit into
chaksaray wants to merge 1 commit into
Conversation
Four mappings from AVE's behavioral vulnerability taxonomy to OWASP Agentic Top 10 entries, each verified against the primary-source PDF's own category text rather than category-name resemblance. Submitted via the manual PR path rather than the JSON classifier intake, matching how MITRE ATLAS entered this crosswalk, since AVE describes attacker behavior rather than defensive controls and doesn't fit the submission schema's controls-catalog shape. - crosswalk/agentic-top10/Agentic_AVE.md: new mapping file, follows the Agentic_AIUC1.md reference-implementation structure (per CONTRIBUTING.md), covering ASI02/ASI04/ASI05 for AVE-2026-00054/00062/00068/00074. Omits the Tools table (AVE's own hard rules explicitly exclude vendor/tool recommendations) and the Implementation-priority table (not meaningful at 3-of-10-entry pilot scale) rather than filling either with invented content, and includes an explicit methodology note on two candidate records excluded during selection for a real ambiguity found against the ASI primary source. - crosswalk/data/entries/ASI02.json, ASI04.json, ASI05.json: appended the corresponding AVE mapping objects to keep the machine-readable layer in sync, per CONTRIBUTING.md; validated against crosswalk/data/schema.json. - crosswalk/README.md: mapping-files badge 67 -> 68. Ran crosswalk's own `node scripts/validate.js`: fixed the 3 errors this change introduced (stale badge count, missing '## Why' heading, missing an ASI01 section — added one explaining ASI01 is deliberately out of scope for this pilot). The remaining ~51 errors/67 warnings reported by that run are pre-existing across other files in the repo, unrelated to this change. Co-Authored-By: Claude Sonnet 5 <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.
AVE is an open behavioral vulnerability taxonomy for agentic AI components (skill files, MCP servers, agent plugins), 80 records with stable immutable IDs, Apache 2.0. github.com/aveproject/ave
Submitting four mappings as a deliberate pilot rather than the full corpus, the clearest and most unambiguous correspondences, so the format and approach can be confirmed before more follow. Each mapping was verified against the OWASP Agentic Top 10 primary-source text directly, not by category-name resemblance, and two additional candidates were excluded during selection when their correspondence turned out genuinely ambiguous (documented in the file's own Methodology note, and in an explicit
### ASI01 — not covered in this pilotsection).Using the manual PR path rather than the JSON submission form: AVE describes attacker-observable behavior rather than defensive controls, so it doesn't fit the submission schema's category enum (ai-governance, regulation, infosec, appsec, threat-model) or its controls-shaped examples. MITRE ATLAS's own entry in this crosswalk came through the same manual path for what looks like the same structural reason.
What's in this PR:
crosswalk/agentic-top10/Agentic_AVE.md— new mapping file, structured to matchAgentic_AIUC1.md(the file CONTRIBUTING.md names as the reference implementation), covering ASI02/ASI04/ASI05.crosswalk/data/entries/ASI02.json,ASI04.json,ASI05.json— appended the corresponding AVE mapping objects; validated againstcrosswalk/data/schema.json.crosswalk/README.md— mapping-files badge 67 → 68 (theframeworks/controlsbadges may also need a bump; left to whatever your generate script computes rather than hand-edited by me without knowing its exact counting logic).Two sections are intentionally not filled in rather than padded with invented content: Tools (AVE's own hard rules explicitly exclude vendor/tool recommendations — it's a vendor-neutral behavioral taxonomy) and Implementation priority (not meaningful yet at 3-of-10-entry pilot scale).
Ran
node scripts/validate.jsbefore opening this: it originally flagged 3 issues caused by this change (stale mapping-files badge, missing## Whyheading, missing an ASI01 section) — all three are fixed in this PR. The remaining ~51 errors / 67 warnings that run reports are pre-existing across other files in the repo and unrelated to this change (confirmed by diffing validator output before/after).Happy to expand, reshape, or reformat as useful.