This canonical paper page includes Commercialization Proof and Related Resources.
ARXIV:2603.28013 · LLM SECURITY · SUBMITTED 31 MAR · 20:21 UTC · FRESHNESS STALE
ARXIV:2603.28013LLM SECURITYSUBMITTED 31 MAR · 20:21 UTCFRESHNESS STALEHaochuan Kevin Wang · arXiv
This research analyzes prompt injection attacks on LLM agents by tracking cryptographic tokens through attack stages to identify defense weaknesses.
Opportunity summary
Pain This research analyzes prompt injection attacks on LLM agents by tracking cryptographic tokens through attack stages to identify defense weaknesses.
Evidence 9 refs | 3 sources | 50% coverage
Blocker Evidence unverified
This research analyzes prompt injection attacks on LLM agents by tracking cryptographic tokens through attack stages to identify defense weaknesses. Prior work measures task-level attack success rate (ASR); we localize the pipeline stage at…
We present a stage-decomposed analysis of prompt injection attacks against five frontier LLM agents. Prior work measures task-level attack success rate (ASR); we localize the pipeline stage at which each model's defense activates.
ScienceToStartup currently rates this 4.0/10 on the public viability pass. Concretely: (1) in our evaluation, exposure is 100% for all five models -- the safety gap is entirely downstream; (2) Claude strips injections at…
LLM Security moved forward this cycle; last verified April 2026. Public score 4.0/10.
Continue into Read for claims, analysis, references, and neighboring papers.
mobile layout uses overflow-hidden min-w-0 break-wordsOpportunity summary
Score4.0Analysis summary
This research analyzes prompt injection attacks on LLM agents by tracking cryptographic tokens through attack stages to identify defense weaknesses.
Loading BUILD…
Paper Pack
10.48550/arXiv.2603.28013This research analyzes prompt injection attacks on LLM agents by tracking cryptographic tokens through attack stages to identify defense weaknesses.
Abstract
We present a stage-decomposed analysis of prompt injection attacks against five frontier LLM agents. Prior work measures task-level attack success rate (ASR); we localize the pipeline stage at which each model's defense activates. We instrument every run with a cryptographic canary token (SECRET-[A-F0-9]{8}) tracked through four kill-chain stages -- Exposed, Persisted, Relayed, Executed -- across four attack surfaces and five defense conditions (764 total runs, 428 no-defense attacked). Our central finding is that model safety is determined not by whether adversarial content is seen, but by whether it is propagated across pipeline stages. Concretely: (1) in our evaluation, exposure is 100% for all five models -- the safety gap is entirely downstream; (2) Claude strips injections at write_memory summarization (0/164 ASR), while GPT-4o-mini propagates canaries without loss (53% ASR, 95% CI: 41--65%); (3) DeepSeek exhibits 0% ASR on memory surfaces and 100% ASR on tool-stream surfaces from the same model -- a complete reversal across injection channels; (4) all four active defense conditions (write_filter, pi_detector, spotlighting, and their combination) produce 100% ASR due to threat-model surface mismatch; (5) a Claude relay node decontaminates downstream agents -- 0/40 canaries survived into shared memory.
Source availability
PDF linkedThe paper record includes a public PDF URL.
Extraction status
Parse run linkedA document parse run is attached to this paper.
Proof status
unverified9 refs; 3 sources; 50% coverage.
What was readable
Derived fallback: Estimated from adjacent evidence; not verified from source.
Viability
Time to MVP
Commercial
Export
Preparing verified analysis
Dimensions overall score 4.0
PROBLEM
This research analyzes prompt injection attacks on LLM agents by tracking cryptographic tokens through attack stages to identify defense weaknesses. Prior work measures task-level attack success rate (ASR); we localize the pipeline stage at which each model's defense activates.
METHOD
We present a stage-decomposed analysis of prompt injection attacks against five frontier LLM agents. Prior work measures task-level attack success rate (ASR); we localize the pipeline stage at which each model's defense activates.
RESULT
ScienceToStartup currently rates this 4.0/10 on the public viability pass. Concretely: (1) in our evaluation, exposure is 100% for all five models -- the safety gap is entirely downstream; (2) Claude strips injections at write_memory summarization (0/164 ASR), while GPT-4o-mini...
WHY NOW
LLM Security moved forward this cycle; last verified April 2026. Public score 4.0/10.
model safety is determined not by whether adversarial content is seen, but by whether it is propagated across pipeline stages
Explicitly stated as the central finding in the abstract, supported by stage-level analysis.
partial
This demonstrates that single-surface evaluation produces a complete mischaracterization of actual safety posture.
Strongly implied by DeepSeek results showing 0% vs 100% ASR across surfaces, with explicit commentary about mischaracterization.
partial
in our evaluation, exposure is 100% for all five models
Explicitly stated multiple times in the abstract and results section with clear numeric evidence.
verified
Claude strips injections at write_memory summarization (0/164 ASR)
Directly stated in the abstract with specific stage localization and numeric result (0/164 ASR).
partial
GPT-4o-mini propagates canaries without loss (53% ASR, 95% CI: 41–65%)
Explicitly stated in abstract with precise percentage and confidence interval.
partial
DeepSeek exhibits 0% ASR on memory surfaces and 100% ASR on tool-stream surfaces from the same model
Directly stated in abstract and results section with specific numeric evidence (0/24 vs 8/8).
partial
all four active defense conditions (write_filter, pi_detector, spotlighting, and their combination) produce 100% ASR due to threat-model surface mismatch
Explicitly stated in abstract but requires checking context about defense conditions; strongly supported by results.
partial
a Claude relay node decontaminates downstream agents -- 0/40 canaries survived into shared memory
Directly stated in abstract with specific numeric evidence (0/40).
partial
Paper-native neighborhood for concepts, methods, materials, markets, and competitors. Missing lanes stay labeled instead of disappearing behind commercialization gates.
Concepts
Methods
Materials
Markets
Competitors
This research analyzes prompt injection attacks on LLM agents by tracking cryptographic tokens through attack stages to identify defense weaknesses.
Segment
LLM Security
Adoption evidence
No public code link in the paper record yet
Commercial read
4.0/10 public viability
Direct
Adjacent
Substitute
Unknown
No indexed public discussion is attached to 2603.28013 yet. That is a visibility signal, not a blank module: the monitor is watching the public channels below.
Hacker News
Not indexed yet
Not indexed yet
Bluesky
Not indexed yet
Preview the source document here, or use the hero PDF action for a new tab.
Reference metadata is not materialized in the public index yet. The source PDF remains the authority; cache refresh is optional.
CITED BY
No citing papers are indexed in the public S2S graph yet. This is an explicit zero-signal state, not a hidden lookup.
Foundation
Extension
Commercially relevant
Owned Distribution
Get the weekly shortlist of commercializable papers, benchmark movers, and proof receipts that matter for product execution.
3/3 checks · 100%
Build Passport
Build passport pending - Proof Lab budget No verified cost estimate / $7.00 cap
status
missing
reason
passport_row_missing
proof status
unverified
cost/budget
No verified cost estimate
confidence low
next verification path
Build brief missing until Build Passport data exists.
Source missing: Build Passport payload.
Experiment plan missing until prototype path is available.
No prototype path attached.
Validation checklist missing until required assets, cost, and regulatory flags are verified.
No checklist artifact is attached to the Build Passport payload.
Derived signals show verified:false until source-backed receipts exist.
Evidence coverage
OpportunityKernel evidence_receipt
9 refs / 3 sources / 50% coverage
stale
Verify missing sources before using this as buyer proof. verified:false
Build readiness
BuildPassport EvidenceState
passport absent
stale
Run Proof Lab or inspect typed missing state. verified:false
Artifact maturity
GitHub and Hugging Face maturity payloads
No public artifact surface observed
stale
Open source artifacts or mark the gap as missing. verified:false
Technical feasibility
partial
Current read
Runnable path is not fully verified.
Evidence
No Build Passport payload attached.
Gaps
Next test
Run minimal reproduction from the Build Passport prototype path.
Market urgency
partial
Current read
Research evidence exists; buyer urgency still needs source proof.
Evidence
9 references, 3 sources, 50% evidence coverage.
Gaps
Next test
Collect buyer interview, deployment evidence, or cited demand signal.
Buyer clarity
missing
Current read
No budget owner is verified for this paper.
Evidence
Build tab has no CRM, procurement, or operator source.
Gaps
Next test
Map target operator, economic buyer, and procurement trigger.
Defensibility
missing
Current read
Defensibility signals are missing.
Evidence
No defensibility receipt attached.
Gaps
Next test
Refresh defensibility bars with source receipts.
Integration burden
missing
Current read
No public implementation surface observed.
Evidence
No GitHub or Hugging Face payload attached.
Gaps
Next test
Write integration checklist from prototype path and target workflow.
Capital intensity
missing
Current read
No observed cost estimate is verified.
Evidence
Cost passport has no observed_usd value.
Gaps
Next test
Run cost passport or mark the cost field not applicable.
Regulatory load
missing
Current read
No regulatory classification is attached.
Evidence
Build Passport ledger does not include regulatory flags.
Gaps
Next test
Classify regulatory flags before commercialization planning.
No named scientific founder assigned.
Paper authors are not treated as operators without consent.
People
No named person assigned.
Gaps
Next verification path
Prototype owner missing.
Build Passport does not name an implementer.
People
No named person assigned.
Gaps
Next verification path
Operator workflow not sourced.
No buyer or workflow interview attached.
People
No named person assigned.
Gaps
Next verification path
No GTM owner verified.
No CRM or outreach source attached.
People
No named person assigned.
Gaps
Next verification path
Regulatory need unclassified.
No clinical or regulatory source attached.
People
No named person assigned.
Gaps
Next verification path
ARTIFACTS
No public artifacts yet.
DEFENSIBILITY
Defensibility and confidence evidence pending.
WATCHTOWER
No verified watchtower monitor rows yet.
FORESIGHT
No prediction yet — minted on next Foresight batch.
OPPORTUNITYKERNEL CHANGES SINCE LAST VIEW
No verified OpportunityKernel changes since the last view.
COMPETITIVE LANDSCAPE UPDATES
No verified competitive landscape changes yet.
RELATED PAPER UPDATES
No verified related paper changes yet.
SIGNAL CANVAS HISTORY AND DELTAS
No Signal Canvas history deltas yet.
TIMELINE
Save this paper to start tracking momentum - commits, demos, and score changes appear here.
No tracked events yet.
Score trend will appear after multiple data points.
BUZZ
Buzz trend pending.