{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp09_archetype_breadth150_20260804","research_id":"eoa_inverse_innovation_exp09_light_prior_art_20260804","cell_id":"substrate_lineage_risk_audit__computer_science","search_lanes":{"direct_problem_and_intervention":{"queries":["generated parser provenance security audit parser generator templates runtime reproducible build","untrusted input parser denial of service resource exhaustion error recovery parser generator","code analysis service parser untrusted source denial of service language server parser vulnerability"],"source_ids":["SRC1","SRC2","SRC3","SRC4"],"no_result_note":null},"synonyms_and_historical_terms":{"queries":["generated code security review provenance generator version templates runtime dependency","software supply chain lineage build inputs generated artifact","parser differential testing regenerated parser AST equivalence fuzzing","language server fuzzing untrusted source editor operations"],"source_ids":["SRC1","SRC2","SRC3"],"no_result_note":null},"products_practices_and_standards":{"queries":["SLSA provenance official build definition external parameters resolved dependencies artifacts","site:tree-sitter.github.io fuzzing generated parser corpus timeout parser official docs","site:owasp.org untrusted input parser resource exhaustion recursion limits","NIST SP 800-218 SSDF generated code compiler tools security review official PDF"],"source_ids":["SRC2","SRC3","SRC4"],"no_result_note":null},"component_combination":{"queries":["parser differential testing regenerated parser AST equivalence fuzzing","parser generator vulnerability denial of service malformed input generated parser CVE","generated parser provenance security audit parser generator templates runtime reproducible build","untrusted input parser denial of service resource exhaustion error recovery parser generator"],"source_ids":["SRC1","SRC2","SRC3","SRC4"],"no_result_note":null}},"sources":[{"source_id":"SRC1","title":"LSPFuzz: Hunting Bugs in Language Servers","publisher":"arXiv / authors from HKUST and collaborating institutions","url":"https://arxiv.org/abs/2510.00532","source_type":"PRIMARY_RESEARCH","claims_supported":["Language servers can crash or expose vulnerabilities while processing untrusted source code.","Testing language services benefits from combining syntax-aware source mutations with editor-operation mutations.","Evaluation on four language servers found 51 bugs, of which 42 were confirmed, 26 fixed, and two assigned CVEs."]},{"source_id":"SRC2","title":"SLSA Provenance","publisher":"Supply-chain Levels for Software Artifacts (SLSA)","url":"https://slsa.dev/spec/v1.0/provenance","source_type":"OFFICIAL_STANDARD","claims_supported":["Artifact provenance can record the build definition, builder, parameters, dependencies, output artifacts, and digests.","Provenance is intended to support verification and rebuilding of an artifact.","Recursive supply-chain analysis should record artifacts that could affect a build, while completeness remains best effort even at higher SLSA levels."]},{"source_id":"SRC3","title":"tree-sitter generate","publisher":"Tree-sitter","url":"https://tree-sitter.github.io/tree-sitter/cli/generate.html","source_type":"FIRST_PARTY_PRODUCT","claims_supported":["A parser generator transforms a structured grammar into generated parser source and supporting files.","Tree-sitter's generated artifacts include parser logic, headers, allocation and array support, grammar representation, and node-type information.","Generation logs expose recovery-state tokens and state transformations, while ABI and runtime choices are explicit generation variables.","A structured grammar artifact can be used to regenerate missing parser source."]},{"source_id":"SRC4","title":"API4:2023 Unrestricted Resource Consumption","publisher":"OWASP Foundation","url":"https://owasp.org/API-Security/editions/2023/en/0xa4-unrestricted-resource-consumption/","source_type":"OFFICIAL_GUIDANCE","claims_supported":["APIs without appropriate execution-time, memory, process, operation-count, and payload-size limits are exposed to resource-consumption attacks and denial of service.","Recommended containment includes CPU and memory limits, process isolation mechanisms, payload bounds, validation, and throttling.","Traditional rate limiting alone may not prevent a single request or batch from exhausting server resources."]}],"problem_evidence":{"status":"PARTLY_SUPPORTED","finding":"The core operational problem is visible: primary research shows that language servers processing untrusted source and editor operations have real crash and vulnerability exposure; Tree-sitter documentation confirms that grammar, generator options, recovery configuration, ABI, support headers, and runtime choices enter generated parser artifacts; and OWASP recognizes missing per-request CPU, memory, time, and operation limits as a denial-of-service condition. The retained sources do not establish how often reviewers exclude generated parser lineages, nor that unidentified generator provenance is prevalent, so that audit-blind-spot portion remains a proposal-specific hypothesis.","source_ids":["SRC1","SRC3","SRC4"]},"closest_prior_art":[{"name":"LSPFuzz language-server testing","source_ids":["SRC1"],"overlap":"Exercises source text together with editor operations and measures real language-server failures in the untrusted, incrementally edited context described by the proposal.","remaining_difference":"It is a bug-finding technique, not a governance method that reconstructs grammar-to-artifact lineage, compares origin assumptions, tests declared regeneration, and requires an owned substrate disposition."},{"name":"SLSA build provenance applied to parser generation","source_ids":["SRC2"],"overlap":"Captures build definitions, parameters, dependencies, builders, artifacts, and digests sufficient to verify or attempt reproduction of a generated parser.","remaining_difference":"SLSA addresses build provenance and supply-chain verification generally; it does not prescribe parser-specific context-shift probes, syntax-tree equivalence checks, origin-context analysis, containment, or a parser release gate."},{"name":"Tree-sitter generation and regeneration workflow","source_ids":["SRC3"],"overlap":"Makes the grammar-to-generated-parser transformation inspectable and identifies recovery state, ABI, runtime, parser source, support headers, and structured grammar artifacts relevant to reproducibility.","remaining_difference":"The product workflow does not itself establish an inherited-risk audit, compare trusted batch assumptions with a hostile multi-tenant service, assign service-risk ownership, or condition deployment on the combined evidence."},{"name":"OWASP unrestricted-resource-consumption controls","source_ids":["SRC4"],"overlap":"Recommends CPU, memory, time, process, request-size, and operation limits, including isolation mechanisms, matching the proposal's request-budget and isolated-worker containment options.","remaining_difference":"The guidance is generic API containment and does not reconstruct parser lineage or test whether generator, grammar, templates, runtime, recovery, or normalization explain resource or syntax-tree discrepancies."}],"prior_art_disposition":"ADJACENT_PRIOR_ART","contrastive_claim_remaining":"For a copied generated parser newly exposed through a multi-tenant code-analysis service, jointly treating the grammar, generator executable and options, templates, generated source, and support runtime as one governed substrate will produce decision-relevant evidence beyond ordinary language-server fuzzing, build provenance, regeneration tooling, or generic request limits. Specifically, lineage reconstruction plus a controlled regeneration and capped context-shift corpus should either identify a material reproducibility or behavioral discrepancy attributable to a lineage variable, or provide evidence supporting an explicit owned disposition and release gate.","contrastive_claim_falsifier":"The contrastive claim is falsified if complete lineage reconstruction yields byte-for-byte reproducibility, declared lineage variables do not explain or stratify termination, resource use, recovery behavior, normalization behavior, or syntax-tree output, and adding the lineage-and-origin-context review changes neither controls nor the deployment decision compared with ordinary fuzzing, dependency provenance, and existing request limits.","gates":{"adequate_source_search":{"status":"PASS","rationale":"The bounded search covered direct wording, older and adjacent terminology, standards and first-party tooling, operational practices, vulnerabilities, fuzzing, regeneration, provenance, and resource containment. Exactly four opened sources from four publishers were retained, including primary research, an official specification, first-party documentation, and official guidance.","source_ids":["SRC1","SRC2","SRC3","SRC4"]},"supported_problem":{"status":"PASS","rationale":"The problem is partly supported: real language-server bugs under untrusted source and edit operations, generator-to-parser transformation details, and API resource-exhaustion risks are documented. Prevalence of the proposed review-boundary blind spot remains unestablished.","source_ids":["SRC1","SRC3","SRC4"]},"distinct_testable_claim":{"status":"PASS","rationale":"The remaining claim is narrower than the adjacent practices and is falsifiable by measuring reproducibility and whether lineage variables explain behavioral differences or alter an owned deployment disposition.","source_ids":["SRC1","SRC2","SRC3","SRC4"]},"bounded_next_test":{"status":"PASS","rationale":"One parser, five working days, a fixed artifact inventory, one regeneration attempt, at most five context differences, a predeclared synthetic corpus, fixed resource caps, and recorded termination, resource, and tree-equivalence outcomes form a bounded test. The source set provides concrete provenance fields, generation artifacts, language-server test inputs, and resource limits to operationalize it.","source_ids":["SRC1","SRC2","SRC3","SRC4"]},"no_obvious_safety_or_authority_stop":{"status":"PASS","rationale":"The authorized first step is non-production, synthetic, isolated, resource-capped, reversible, and leaves deployment decisions with named product, security, and reliability authorities. It avoids production traffic and external disclosure; OWASP guidance supports isolation and resource caps, while SLSA supplies a non-destructive provenance model.","source_ids":["SRC2","SRC4"]}},"screen_survival":true,"world_novelty_boundary":"This bounded public-web screen found adjacent prior art for every major component but no retained source prescribing the full parser-specific combination of lineage reconstruction, origin-context comparison, controlled regeneration, capped behavioral and syntax-tree comparison, interim containment, owned discrepancy disposition, and release gating. That result supports only coarse researchability and a remaining contrastive hypothesis; it does not establish world novelty, patentability, market size, expert acceptance, prevalence, or realized value."}