DESERVE: A Framework for Detecting Program Security Vulnerability Exploitations
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
It is difficult to develop a program that is completely free from vulnerabilities. Despite the application of many approaches to secure programs, vulnerability exploitations occur in real-world in large numbers. Exploitations of vulnerabilities may corrupt memory spaces and program states, lead to denial of services and authorization bypassing, and leak sensitive information. Monitoring at the program code level can be a way of vulnerability exploitation detection at runtime. In this work, we propose a monitor embedding framework DESERVE (a framework for Detecting program Security Vulnerability Exploitations). DESERVE identifies exploitable statements from source code based on static backward slicing and embeds necessary code to detect attacks. During the deployment stage, the enhanced programs execute exploitable statements in a separate test environment. Unlike traditional monitors that extract and store program state information to compare with vulnerable free program states to detect exploitation, our approach does not need to save state information. Moreover, the slicing technique allows us avoid the tracking of fine grained level of information about runtime program environments such as input flow and memory state. We implement DESERVE for detecting buffer overflow, SQL injection, and cross-site scripting attacks. We evaluate our approach for real-world programs implemented in C and PHP languages. The results show that the approach can detect some of the well-known attacks. Moreover, the approach imposes negligible runtime overhead.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it