feat(rate-limiting): Implement core rate limiting functionality with configuration, decision-making, metrics, middleware, and service registration
- Add RateLimitConfig for configuration management with YAML binding support. - Introduce RateLimitDecision to encapsulate the result of rate limit checks. - Implement RateLimitMetrics for OpenTelemetry metrics tracking. - Create RateLimitMiddleware for enforcing rate limits on incoming requests. - Develop RateLimitService to orchestrate instance and environment rate limit checks. - Add RateLimitServiceCollectionExtensions for dependency injection registration.
This commit is contained in:
master
143
datasets/reachability/corpus.json
Normal file
143
datasets/reachability/corpus.json
Normal file
@@ -0,0 +1,143 @@
|
||||
{
|
||||
"$schema": "https://stellaops.io/schemas/corpus-index.v1.json",
|
||||
"version": "1.0.0",
|
||||
"description": "Ground-truth corpus for binary reachability benchmarking",
|
||||
"createdAt": "2025-12-17T00:00:00Z",
|
||||
"samples": [
|
||||
{
|
||||
"sampleId": "gt-0001",
|
||||
"category": "basic",
|
||||
"path": "ground-truth/basic/gt-0001/sample.manifest.json",
|
||||
"description": "Direct call to vulnerable sink from main"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0002",
|
||||
"category": "basic",
|
||||
"path": "ground-truth/basic/gt-0002/sample.manifest.json",
|
||||
"description": "Two-hop call chain to vulnerable sink"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0003",
|
||||
"category": "basic",
|
||||
"path": "ground-truth/basic/gt-0003/sample.manifest.json",
|
||||
"description": "Three-hop call chain with multiple sinks"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0004",
|
||||
"category": "basic",
|
||||
"path": "ground-truth/basic/gt-0004/sample.manifest.json",
|
||||
"description": "Function pointer call to sink"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0005",
|
||||
"category": "basic",
|
||||
"path": "ground-truth/basic/gt-0005/sample.manifest.json",
|
||||
"description": "Recursive function with sink"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0006",
|
||||
"category": "indirect",
|
||||
"path": "ground-truth/indirect/gt-0006/sample.manifest.json",
|
||||
"description": "Indirect call via callback"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0007",
|
||||
"category": "indirect",
|
||||
"path": "ground-truth/indirect/gt-0007/sample.manifest.json",
|
||||
"description": "Virtual function dispatch"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0008",
|
||||
"category": "guarded",
|
||||
"path": "ground-truth/guarded/gt-0008/sample.manifest.json",
|
||||
"description": "Sink behind constant false guard"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0009",
|
||||
"category": "guarded",
|
||||
"path": "ground-truth/guarded/gt-0009/sample.manifest.json",
|
||||
"description": "Sink behind input-dependent guard"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0010",
|
||||
"category": "guarded",
|
||||
"path": "ground-truth/guarded/gt-0010/sample.manifest.json",
|
||||
"description": "Sink behind environment variable guard"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0011",
|
||||
"category": "basic",
|
||||
"path": "ground-truth/basic/gt-0011/sample.manifest.json",
|
||||
"description": "Unreachable sink - dead code after return"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0012",
|
||||
"category": "basic",
|
||||
"path": "ground-truth/basic/gt-0012/sample.manifest.json",
|
||||
"description": "Unreachable sink - never called function"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0013",
|
||||
"category": "basic",
|
||||
"path": "ground-truth/basic/gt-0013/sample.manifest.json",
|
||||
"description": "Unreachable sink - #ifdef disabled"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0014",
|
||||
"category": "guarded",
|
||||
"path": "ground-truth/guarded/gt-0014/sample.manifest.json",
|
||||
"description": "Unreachable sink - constant true early return"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0015",
|
||||
"category": "guarded",
|
||||
"path": "ground-truth/guarded/gt-0015/sample.manifest.json",
|
||||
"description": "Unreachable sink - impossible branch condition"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0016",
|
||||
"category": "stripped",
|
||||
"path": "ground-truth/stripped/gt-0016/sample.manifest.json",
|
||||
"description": "Stripped binary - reachable sink"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0017",
|
||||
"category": "stripped",
|
||||
"path": "ground-truth/stripped/gt-0017/sample.manifest.json",
|
||||
"description": "Stripped binary - unreachable sink"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0018",
|
||||
"category": "obfuscated",
|
||||
"path": "ground-truth/obfuscated/gt-0018/sample.manifest.json",
|
||||
"description": "Control flow obfuscation - reachable"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0019",
|
||||
"category": "obfuscated",
|
||||
"path": "ground-truth/obfuscated/gt-0019/sample.manifest.json",
|
||||
"description": "String obfuscation - reachable"
|
||||
},
|
||||
{
|
||||
"sampleId": "gt-0020",
|
||||
"category": "callback",
|
||||
"path": "ground-truth/callback/gt-0020/sample.manifest.json",
|
||||
"description": "Async callback chain - reachable"
|
||||
}
|
||||
],
|
||||
"statistics": {
|
||||
"totalSamples": 20,
|
||||
"byCategory": {
|
||||
"basic": 8,
|
||||
"indirect": 2,
|
||||
"guarded": 4,
|
||||
"stripped": 2,
|
||||
"obfuscated": 2,
|
||||
"callback": 2
|
||||
},
|
||||
"byExpected": {
|
||||
"reachable": 13,
|
||||
"unreachable": 7
|
||||
}
|
||||
}
|
||||
}
|
||||
18
datasets/reachability/ground-truth/basic/gt-0001/main.c
Normal file
18
datasets/reachability/ground-truth/basic/gt-0001/main.c
Normal file
@@ -0,0 +1,18 @@
|
||||
// gt-0001: Direct call to vulnerable sink from main
|
||||
// Expected: REACHABLE (tier: executed)
|
||||
// Vulnerability: CWE-120 (Buffer Copy without Checking Size)
|
||||
|
||||
#include <stdio.h>
|
||||
#include <string.h>
|
||||
|
||||
int main(int argc, char *argv[]) {
|
||||
char buffer[32];
|
||||
|
||||
if (argc > 1) {
|
||||
// Vulnerable: strcpy without bounds checking
|
||||
strcpy(buffer, argv[1]); // SINK: CWE-120
|
||||
printf("Input: %s\n", buffer);
|
||||
}
|
||||
|
||||
return 0;
|
||||
}
|
||||
@@ -0,0 +1,29 @@
|
||||
{
|
||||
"$schema": "https://stellaops.io/schemas/sample-manifest.v1.json",
|
||||
"sampleId": "gt-0001",
|
||||
"version": "1.0.0",
|
||||
"category": "basic",
|
||||
"description": "Direct call to vulnerable sink from main - REACHABLE",
|
||||
"language": "c",
|
||||
"expectedResult": {
|
||||
"reachable": true,
|
||||
"tier": "executed",
|
||||
"confidence": 1.0
|
||||
},
|
||||
"source": {
|
||||
"files": ["main.c"],
|
||||
"entrypoint": "main",
|
||||
"sink": "strcpy",
|
||||
"vulnerability": "CWE-120"
|
||||
},
|
||||
"callChain": [
|
||||
{"function": "main", "file": "main.c", "line": 5},
|
||||
{"function": "strcpy", "file": "<libc>", "line": null}
|
||||
],
|
||||
"annotations": {
|
||||
"notes": "Simplest reachable case - direct call from entrypoint to vulnerable function",
|
||||
"difficulty": "trivial"
|
||||
},
|
||||
"createdAt": "2025-12-17T00:00:00Z",
|
||||
"createdBy": "corpus-team"
|
||||
}
|
||||
22
datasets/reachability/ground-truth/basic/gt-0002/main.c
Normal file
22
datasets/reachability/ground-truth/basic/gt-0002/main.c
Normal file
@@ -0,0 +1,22 @@
|
||||
// gt-0002: Two-hop call chain to vulnerable sink
|
||||
// Expected: REACHABLE (tier: executed)
|
||||
// Vulnerability: CWE-134 (Format String)
|
||||
|
||||
#include <stdio.h>
|
||||
#include <string.h>
|
||||
|
||||
void format_message(const char *user_input, char *output) {
|
||||
// Vulnerable: format string from user input
|
||||
sprintf(output, user_input); // SINK: CWE-134
|
||||
}
|
||||
|
||||
int main(int argc, char *argv[]) {
|
||||
char buffer[256];
|
||||
|
||||
if (argc > 1) {
|
||||
format_message(argv[1], buffer);
|
||||
printf("Result: %s\n", buffer);
|
||||
}
|
||||
|
||||
return 0;
|
||||
}
|
||||
@@ -0,0 +1,30 @@
|
||||
{
|
||||
"$schema": "https://stellaops.io/schemas/sample-manifest.v1.json",
|
||||
"sampleId": "gt-0002",
|
||||
"version": "1.0.0",
|
||||
"category": "basic",
|
||||
"description": "Two-hop call chain to vulnerable sink - REACHABLE",
|
||||
"language": "c",
|
||||
"expectedResult": {
|
||||
"reachable": true,
|
||||
"tier": "executed",
|
||||
"confidence": 1.0
|
||||
},
|
||||
"source": {
|
||||
"files": ["main.c"],
|
||||
"entrypoint": "main",
|
||||
"sink": "sprintf",
|
||||
"vulnerability": "CWE-134"
|
||||
},
|
||||
"callChain": [
|
||||
{"function": "main", "file": "main.c", "line": 15},
|
||||
{"function": "format_message", "file": "main.c", "line": 7},
|
||||
{"function": "sprintf", "file": "<libc>", "line": null}
|
||||
],
|
||||
"annotations": {
|
||||
"notes": "Two-hop chain: main -> helper -> sink",
|
||||
"difficulty": "easy"
|
||||
},
|
||||
"createdAt": "2025-12-17T00:00:00Z",
|
||||
"createdBy": "corpus-team"
|
||||
}
|
||||
25
datasets/reachability/ground-truth/basic/gt-0003/main.c
Normal file
25
datasets/reachability/ground-truth/basic/gt-0003/main.c
Normal file
@@ -0,0 +1,25 @@
|
||||
// gt-0003: Three-hop call chain with command injection
|
||||
// Expected: REACHABLE (tier: executed)
|
||||
// Vulnerability: CWE-78 (OS Command Injection)
|
||||
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
|
||||
void execute_command(const char *cmd) {
|
||||
// Vulnerable: system call with user input
|
||||
system(cmd); // SINK: CWE-78
|
||||
}
|
||||
|
||||
void process_input(const char *input) {
|
||||
char command[256];
|
||||
snprintf(command, sizeof(command), "echo %s", input);
|
||||
execute_command(command);
|
||||
}
|
||||
|
||||
int main(int argc, char *argv[]) {
|
||||
if (argc > 1) {
|
||||
process_input(argv[1]);
|
||||
}
|
||||
return 0;
|
||||
}
|
||||
@@ -0,0 +1,31 @@
|
||||
{
|
||||
"$schema": "https://stellaops.io/schemas/sample-manifest.v1.json",
|
||||
"sampleId": "gt-0003",
|
||||
"version": "1.0.0",
|
||||
"category": "basic",
|
||||
"description": "Three-hop call chain with multiple sinks - REACHABLE",
|
||||
"language": "c",
|
||||
"expectedResult": {
|
||||
"reachable": true,
|
||||
"tier": "executed",
|
||||
"confidence": 1.0
|
||||
},
|
||||
"source": {
|
||||
"files": ["main.c"],
|
||||
"entrypoint": "main",
|
||||
"sink": "system",
|
||||
"vulnerability": "CWE-78"
|
||||
},
|
||||
"callChain": [
|
||||
{"function": "main", "file": "main.c", "line": 20},
|
||||
{"function": "process_input", "file": "main.c", "line": 12},
|
||||
{"function": "execute_command", "file": "main.c", "line": 6},
|
||||
{"function": "system", "file": "<libc>", "line": null}
|
||||
],
|
||||
"annotations": {
|
||||
"notes": "Three-hop chain demonstrating command injection path",
|
||||
"difficulty": "easy"
|
||||
},
|
||||
"createdAt": "2025-12-17T00:00:00Z",
|
||||
"createdBy": "corpus-team"
|
||||
}
|
||||
37
datasets/reachability/ground-truth/basic/gt-0004/main.c
Normal file
37
datasets/reachability/ground-truth/basic/gt-0004/main.c
Normal file
@@ -0,0 +1,37 @@
|
||||
// gt-0004: Function pointer call to sink
|
||||
// Expected: REACHABLE (tier: executed)
|
||||
// Vulnerability: CWE-120 (Buffer Copy without Checking Size)
|
||||
|
||||
#include <stdio.h>
|
||||
#include <string.h>
|
||||
|
||||
typedef void (*copy_func_t)(char *, const char *);
|
||||
|
||||
void copy_data(char *dest, const char *src) {
|
||||
// Vulnerable: strcpy without bounds check
|
||||
strcpy(dest, src); // SINK: CWE-120
|
||||
}
|
||||
|
||||
void safe_copy(char *dest, const char *src) {
|
||||
strncpy(dest, src, 31);
|
||||
dest[31] = '\0';
|
||||
}
|
||||
|
||||
int main(int argc, char *argv[]) {
|
||||
char buffer[32];
|
||||
copy_func_t copier;
|
||||
|
||||
// Function pointer assignment - harder for static analysis
|
||||
if (argc > 2 && argv[2][0] == 's') {
|
||||
copier = safe_copy;
|
||||
} else {
|
||||
copier = copy_data; // Vulnerable path selected
|
||||
}
|
||||
|
||||
if (argc > 1) {
|
||||
copier(buffer, argv[1]); // Indirect call
|
||||
printf("Result: %s\n", buffer);
|
||||
}
|
||||
|
||||
return 0;
|
||||
}
|
||||
@@ -0,0 +1,31 @@
|
||||
{
|
||||
"$schema": "https://stellaops.io/schemas/sample-manifest.v1.json",
|
||||
"sampleId": "gt-0004",
|
||||
"version": "1.0.0",
|
||||
"category": "basic",
|
||||
"description": "Function pointer call to sink - REACHABLE",
|
||||
"language": "c",
|
||||
"expectedResult": {
|
||||
"reachable": true,
|
||||
"tier": "executed",
|
||||
"confidence": 0.9
|
||||
},
|
||||
"source": {
|
||||
"files": ["main.c"],
|
||||
"entrypoint": "main",
|
||||
"sink": "strcpy",
|
||||
"vulnerability": "CWE-120"
|
||||
},
|
||||
"callChain": [
|
||||
{"function": "main", "file": "main.c", "line": 18},
|
||||
{"function": "<function_ptr>", "file": "main.c", "line": 19},
|
||||
{"function": "copy_data", "file": "main.c", "line": 8},
|
||||
{"function": "strcpy", "file": "<libc>", "line": null}
|
||||
],
|
||||
"annotations": {
|
||||
"notes": "Indirect call via function pointer - harder for static analysis",
|
||||
"difficulty": "medium"
|
||||
},
|
||||
"createdAt": "2025-12-17T00:00:00Z",
|
||||
"createdBy": "corpus-team"
|
||||
}
|
||||
31
datasets/reachability/ground-truth/basic/gt-0005/main.c
Normal file
31
datasets/reachability/ground-truth/basic/gt-0005/main.c
Normal file
@@ -0,0 +1,31 @@
|
||||
// gt-0005: Recursive function with sink
|
||||
// Expected: REACHABLE (tier: executed)
|
||||
// Vulnerability: CWE-134 (Format String)
|
||||
|
||||
#include <stdio.h>
|
||||
#include <string.h>
|
||||
|
||||
char result[1024];
|
||||
|
||||
void process_recursive(const char *input, int depth) {
|
||||
if (depth <= 0 || strlen(input) == 0) {
|
||||
return;
|
||||
}
|
||||
|
||||
// Vulnerable: format string in recursive context
|
||||
sprintf(result + strlen(result), input); // SINK: CWE-134
|
||||
|
||||
// Recurse with modified input
|
||||
process_recursive(input + 1, depth - 1);
|
||||
}
|
||||
|
||||
int main(int argc, char *argv[]) {
|
||||
result[0] = '\0';
|
||||
|
||||
if (argc > 1) {
|
||||
process_recursive(argv[1], 5);
|
||||
printf("Result: %s\n", result);
|
||||
}
|
||||
|
||||
return 0;
|
||||
}
|
||||
@@ -0,0 +1,31 @@
|
||||
{
|
||||
"$schema": "https://stellaops.io/schemas/sample-manifest.v1.json",
|
||||
"sampleId": "gt-0005",
|
||||
"version": "1.0.0",
|
||||
"category": "basic",
|
||||
"description": "Recursive function with sink - REACHABLE",
|
||||
"language": "c",
|
||||
"expectedResult": {
|
||||
"reachable": true,
|
||||
"tier": "executed",
|
||||
"confidence": 1.0
|
||||
},
|
||||
"source": {
|
||||
"files": ["main.c"],
|
||||
"entrypoint": "main",
|
||||
"sink": "sprintf",
|
||||
"vulnerability": "CWE-134"
|
||||
},
|
||||
"callChain": [
|
||||
{"function": "main", "file": "main.c", "line": 22},
|
||||
{"function": "process_recursive", "file": "main.c", "line": 14},
|
||||
{"function": "process_recursive", "file": "main.c", "line": 14},
|
||||
{"function": "sprintf", "file": "<libc>", "line": null}
|
||||
],
|
||||
"annotations": {
|
||||
"notes": "Recursive call pattern - tests loop/recursion handling",
|
||||
"difficulty": "medium"
|
||||
},
|
||||
"createdAt": "2025-12-17T00:00:00Z",
|
||||
"createdBy": "corpus-team"
|
||||
}
|
||||
@@ -0,0 +1,25 @@
|
||||
// gt-0011: Dead code - function never called
|
||||
// Expected: UNREACHABLE (tier: imported)
|
||||
// Vulnerability: CWE-120 (Buffer Copy without Checking Size)
|
||||
|
||||
#include <stdio.h>
|
||||
#include <string.h>
|
||||
|
||||
// This function is NEVER called - dead code
|
||||
void vulnerable_function(const char *input) {
|
||||
char buffer[32];
|
||||
strcpy(buffer, input); // SINK: CWE-120 (but unreachable)
|
||||
printf("Value: %s\n", buffer);
|
||||
}
|
||||
|
||||
void safe_function(const char *input) {
|
||||
printf("Safe: %.31s\n", input);
|
||||
}
|
||||
|
||||
int main(int argc, char *argv[]) {
|
||||
if (argc > 1) {
|
||||
// Only safe_function is called
|
||||
safe_function(argv[1]);
|
||||
}
|
||||
return 0;
|
||||
}
|
||||
@@ -0,0 +1,27 @@
|
||||
{
|
||||
"$schema": "https://stellaops.io/schemas/sample-manifest.v1.json",
|
||||
"sampleId": "gt-0011",
|
||||
"version": "1.0.0",
|
||||
"category": "unreachable",
|
||||
"description": "Dead code - function never called - UNREACHABLE",
|
||||
"language": "c",
|
||||
"expectedResult": {
|
||||
"reachable": false,
|
||||
"tier": "imported",
|
||||
"confidence": 1.0
|
||||
},
|
||||
"source": {
|
||||
"files": ["main.c"],
|
||||
"entrypoint": "main",
|
||||
"sink": "strcpy",
|
||||
"vulnerability": "CWE-120"
|
||||
},
|
||||
"callChain": null,
|
||||
"annotations": {
|
||||
"notes": "Vulnerable function exists but is never called from any reachable path",
|
||||
"difficulty": "trivial",
|
||||
"reason": "dead_code"
|
||||
},
|
||||
"createdAt": "2025-12-17T00:00:00Z",
|
||||
"createdBy": "corpus-team"
|
||||
}
|
||||
@@ -0,0 +1,28 @@
|
||||
// gt-0012: Compile-time constant false condition
|
||||
// Expected: UNREACHABLE (tier: imported)
|
||||
// Vulnerability: CWE-120 (Buffer Overflow)
|
||||
|
||||
#include <stdio.h>
|
||||
#include <string.h>
|
||||
|
||||
#define DEBUG_MODE 0 // Compile-time constant
|
||||
|
||||
int main(int argc, char *argv[]) {
|
||||
char buffer[64];
|
||||
|
||||
// This branch is constant false - will be optimized out
|
||||
if (DEBUG_MODE) {
|
||||
// Vulnerable code in dead branch
|
||||
gets(buffer); // SINK: CWE-120 (but unreachable)
|
||||
printf("Debug: %s\n", buffer);
|
||||
} else {
|
||||
// Safe path always taken
|
||||
if (argc > 1) {
|
||||
strncpy(buffer, argv[1], sizeof(buffer) - 1);
|
||||
buffer[sizeof(buffer) - 1] = '\0';
|
||||
printf("Input: %s\n", buffer);
|
||||
}
|
||||
}
|
||||
|
||||
return 0;
|
||||
}
|
||||
@@ -0,0 +1,27 @@
|
||||
{
|
||||
"$schema": "https://stellaops.io/schemas/sample-manifest.v1.json",
|
||||
"sampleId": "gt-0012",
|
||||
"version": "1.0.0",
|
||||
"category": "unreachable",
|
||||
"description": "Compile-time constant false condition - UNREACHABLE",
|
||||
"language": "c",
|
||||
"expectedResult": {
|
||||
"reachable": false,
|
||||
"tier": "imported",
|
||||
"confidence": 1.0
|
||||
},
|
||||
"source": {
|
||||
"files": ["main.c"],
|
||||
"entrypoint": "main",
|
||||
"sink": "gets",
|
||||
"vulnerability": "CWE-120"
|
||||
},
|
||||
"callChain": null,
|
||||
"annotations": {
|
||||
"notes": "Sink is behind a constant false condition that will be optimized out",
|
||||
"difficulty": "easy",
|
||||
"reason": "constant_false"
|
||||
},
|
||||
"createdAt": "2025-12-17T00:00:00Z",
|
||||
"createdBy": "corpus-team"
|
||||
}
|
||||
@@ -0,0 +1,27 @@
|
||||
// gt-0013: Ifdef-excluded code path
|
||||
// Expected: UNREACHABLE (tier: imported)
|
||||
// Vulnerability: CWE-78 (OS Command Injection)
|
||||
// Compile with: gcc -DPRODUCTION main.c (LEGACY_SHELL not defined)
|
||||
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
|
||||
#define PRODUCTION
|
||||
|
||||
void process_command(const char *cmd) {
|
||||
#ifdef LEGACY_SHELL
|
||||
// This code is excluded when LEGACY_SHELL is not defined
|
||||
system(cmd); // SINK: CWE-78 (but unreachable - ifdef excluded)
|
||||
#else
|
||||
// Safe path: just print, don't execute
|
||||
printf("Would execute: %s\n", cmd);
|
||||
#endif
|
||||
}
|
||||
|
||||
int main(int argc, char *argv[]) {
|
||||
if (argc > 1) {
|
||||
process_command(argv[1]);
|
||||
}
|
||||
return 0;
|
||||
}
|
||||
@@ -0,0 +1,27 @@
|
||||
{
|
||||
"$schema": "https://stellaops.io/schemas/sample-manifest.v1.json",
|
||||
"sampleId": "gt-0013",
|
||||
"version": "1.0.0",
|
||||
"category": "unreachable",
|
||||
"description": "Ifdef-excluded code path - UNREACHABLE",
|
||||
"language": "c",
|
||||
"expectedResult": {
|
||||
"reachable": false,
|
||||
"tier": "imported",
|
||||
"confidence": 1.0
|
||||
},
|
||||
"source": {
|
||||
"files": ["main.c"],
|
||||
"entrypoint": "main",
|
||||
"sink": "system",
|
||||
"vulnerability": "CWE-78"
|
||||
},
|
||||
"callChain": null,
|
||||
"annotations": {
|
||||
"notes": "Vulnerable code excluded by preprocessor directive",
|
||||
"difficulty": "easy",
|
||||
"reason": "preprocessor_excluded"
|
||||
},
|
||||
"createdAt": "2025-12-17T00:00:00Z",
|
||||
"createdBy": "corpus-team"
|
||||
}
|
||||
121
datasets/reachability/schemas/corpus-sample.v1.json
Normal file
121
datasets/reachability/schemas/corpus-sample.v1.json
Normal file
@@ -0,0 +1,121 @@
|
||||
{
|
||||
"$schema": "http://json-schema.org/draft-07/schema#",
|
||||
"$id": "https://stellaops.io/schemas/corpus-sample.v1.json",
|
||||
"title": "CorpusSample",
|
||||
"description": "Schema for ground-truth corpus samples used in reachability benchmarking",
|
||||
"type": "object",
|
||||
"required": ["sampleId", "name", "format", "arch", "sinks"],
|
||||
"properties": {
|
||||
"sampleId": {
|
||||
"type": "string",
|
||||
"pattern": "^gt-[0-9]{4}$",
|
||||
"description": "Unique identifier for the sample (e.g., gt-0001)"
|
||||
},
|
||||
"name": {
|
||||
"type": "string",
|
||||
"description": "Human-readable name for the sample"
|
||||
},
|
||||
"description": {
|
||||
"type": "string",
|
||||
"description": "Detailed description of what this sample tests"
|
||||
},
|
||||
"category": {
|
||||
"type": "string",
|
||||
"enum": ["basic", "indirect", "stripped", "obfuscated", "guarded", "callback", "virtual"],
|
||||
"description": "Sample category for organization"
|
||||
},
|
||||
"format": {
|
||||
"type": "string",
|
||||
"enum": ["elf64", "elf32", "pe64", "pe32", "macho64", "macho32"],
|
||||
"description": "Binary format"
|
||||
},
|
||||
"arch": {
|
||||
"type": "string",
|
||||
"enum": ["x86_64", "x86", "aarch64", "arm32", "riscv64"],
|
||||
"description": "Target architecture"
|
||||
},
|
||||
"language": {
|
||||
"type": "string",
|
||||
"enum": ["c", "cpp", "rust", "go"],
|
||||
"description": "Source language (for reference)"
|
||||
},
|
||||
"compiler": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"name": { "type": "string" },
|
||||
"version": { "type": "string" },
|
||||
"flags": { "type": "array", "items": { "type": "string" } }
|
||||
},
|
||||
"description": "Compiler information used to build the sample"
|
||||
},
|
||||
"entryPoint": {
|
||||
"type": "string",
|
||||
"default": "main",
|
||||
"description": "Entry point function name"
|
||||
},
|
||||
"sinks": {
|
||||
"type": "array",
|
||||
"minItems": 1,
|
||||
"items": {
|
||||
"type": "object",
|
||||
"required": ["sinkId", "signature", "expected"],
|
||||
"properties": {
|
||||
"sinkId": {
|
||||
"type": "string",
|
||||
"pattern": "^sink-[0-9]{3}$",
|
||||
"description": "Unique sink identifier within the sample"
|
||||
},
|
||||
"signature": {
|
||||
"type": "string",
|
||||
"description": "Function signature of the sink"
|
||||
},
|
||||
"sinkType": {
|
||||
"type": "string",
|
||||
"enum": ["memory_corruption", "command_injection", "sql_injection", "path_traversal", "format_string", "crypto_weakness", "custom"],
|
||||
"description": "Type of vulnerability represented by the sink"
|
||||
},
|
||||
"expected": {
|
||||
"type": "string",
|
||||
"enum": ["reachable", "unreachable", "conditional"],
|
||||
"description": "Expected reachability determination"
|
||||
},
|
||||
"expectedPaths": {
|
||||
"type": "array",
|
||||
"items": {
|
||||
"type": "array",
|
||||
"items": { "type": "string" }
|
||||
},
|
||||
"description": "Expected call paths from entry to sink (for reachable sinks)"
|
||||
},
|
||||
"guardConditions": {
|
||||
"type": "array",
|
||||
"items": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"variable": { "type": "string" },
|
||||
"condition": { "type": "string" },
|
||||
"value": { "type": "string" }
|
||||
}
|
||||
},
|
||||
"description": "Guard conditions that protect the sink (for conditional sinks)"
|
||||
},
|
||||
"notes": {
|
||||
"type": "string",
|
||||
"description": "Additional notes about this sink"
|
||||
}
|
||||
}
|
||||
},
|
||||
"description": "List of sinks with expected reachability"
|
||||
},
|
||||
"metadata": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"createdAt": { "type": "string", "format": "date-time" },
|
||||
"createdBy": { "type": "string" },
|
||||
"version": { "type": "string" },
|
||||
"sha256": { "type": "string", "pattern": "^[a-f0-9]{64}$" }
|
||||
},
|
||||
"description": "Metadata about the sample"
|
||||
}
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user