CalvinBackup/agentic_security
1
0
Fork 0
mirror of https://github.com/msoedov/agentic_security.git synced 2026-09-29 03:11:51 +02:00
Code Issues Packages Projects Releases Wiki Activity
749 Commits 18 Branches 23 Tags
f43472cc9b6ffa47d4afdbd7afbbee982f8a10f7
Commit Graph
100 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Alexander Myasoedov f43472cc9b fix: ci poetry installs into system python 
poetry.toml sets virtualenvs.create=false, so pipx poetry installed into
the runner's system Python 3.12 and failed uninstalling Debian packages.
Force venv creation in CI via POETRY_VIRTUALENVS_CREATE. Drop the poetry
cache from release, which never creates a venv.
 2026-09-22 16:58:07 +03:00
Alexander Myasoedov bdf2208f6b Merge pull request #346 from jiawenlai1109/fix/audio-temp-leak-on-save-error 
fix(audio_generator): clean up temp file when gTTS save() fails
 2026-09-22 16:53:32 +03:00
Alexander Myasoedov 2340bc2f5a Merge pull request #345 from Anai-Guo/fix/hybrid-classifier-self-annotation 
fix(refusal_classifier): quote self-referential annotation to fix import NameError
 2026-09-11 10:08:39 +03:00
Alexander MyasoedovandClaude Fable 5.1 1f2413e2b0 fix: ci pre-commit poetry env 
Install poetry before setup-python and enable the poetry cache so
setup-python activates the 3.14 interpreter. With virtualenvs.create=false
poetry otherwise installs into the runner's system Python 3.12 and fails
uninstalling Debian dist-packages.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Bk6CFfaN3YJFZ6yztQja7D
 2026-09-07 14:08:49 +03:00
Alexander Myasoedov a49706457c fix: pc 2026-09-07 14:03:57 +03:00
Alexander Myasoedov 5a9e80b9af fix: cosmetic 2026-09-07 14:03:57 +03:00
Alexander Myasoedov 3e6b184c4f fix: template fix 2026-09-07 14:03:57 +03:00
Alexander Myasoedov f286ebf21a fix: docs 2026-09-07 14:03:57 +03:00
Alexander Myasoedov b5aa075927 Merge pull request #340 from rwinkelman/feat/scan-cli-spec-309 
feat(cli): add stateless scan command for agent-invocable runs (#309)
 2026-09-07 13:36:03 +03:00
Alexander Myasoedov 35379d1124 fix: build 2026-09-07 13:35:20 +03:00
Alexander Myasoedov fd43b13bb7 fix: syntax error 2026-09-07 13:24:28 +03:00
Alexander Myasoedov c75df250c7 Merge pull request #343 from TianHengZhuang/main 
docs: document module interface and add ModuleProtocol
 2026-09-07 13:16:03 +03:00
Alexander Myasoedov 80015fd803 Merge pull request #342 from Sabina8205/docs/cli-reference 
docs: add CLI reference and correct configuration filename
 2026-09-07 13:14:19 +03:00
Alexander Myasoedov 73f84b7aa8 Merge pull request #338 from rwinkelman/fix/scan-route-logging-199 
feat: add structured logging to scan API routes (#199)
 2026-09-01 16:16:22 +03:00
Alexander Myasoedov 797180ca3a Merge pull request #336 from rwinkelman/fix/verify-integration-error-status-173 
fix(ui): mark integration verify failure on network errors
 2026-08-31 14:17:34 +03:00
Alexander Myasoedov 1ef131421e feat: move to python 3.14 baseline, patch remaining dependabot alerts 
- requires-python >=3.14, pyupgrade --py314-plus, Dockerfile python:3.14-slim
- test matrix 3.14 only
- fastapi ^0.141.1 + starlette 1.6.0 (unblocks 5 starlette alerts)
- python-multipart ^0.0.31 (4 alerts)
- fix TemplateResponse for starlette 1.x request-first signature
- drop teyit hook: abandoned, incompatible with python 3.14
 2026-08-18 19:14:05 +03:00
Alexander Myasoedov 11024e6fca fix(deps): patch 47 dependabot alerts via lock bump 2026-08-18 19:07:13 +03:00
Alexander Myasoedov 2f128ae16c fix(deps): bump pyarrow to 25.0.1 for cp314 wheels 2026-08-18 19:01:16 +03:00
Alexander Myasoedov 3b1067c317 fix(pc): 2026-08-18 18:49:56 +03:00
Alexander Myasoedov 27c160eaa3 Merge pull request #326 from chiruu12/fix/hybrid-classifier-guards 
fix(hybrid): guard zero total weight and log skipped detectors
 2026-08-18 18:40:07 +03:00
Alexander Myasoedov de16779806 Merge pull request #328 from feiiiiii5/fix-circuit-breaker-half-open 
fix(circuit-breaker): reopen on failure during half-open probation
 2026-08-18 18:39:03 +03:00
Alexander Myasoedov 017ba7c168 Merge pull request #329 from feiiiiii5/fix-p95-percentile-index 
fix(executor): compute p95 latency as a real 95th percentile
 2026-08-18 18:38:25 +03:00
Alexander Myasoedov 43922ced3d Merge pull request #330 from ChrisJr404/code-block-stenography 
Add code_block stenography transform (embed prompt in a docstring)
 2026-08-18 18:37:40 +03:00
Alexander Myasoedov 14e45f3788 Merge pull request #331 from ChrisJr404/feat/owasp-llm-badges 
Add OWASP LLM Top 10 category badges to the scan UI
 2026-08-18 18:36:50 +03:00
Alexander Myasoedov c8458d73c5 Merge pull request #320 from DevamShah/fix-cors-credentials-and-icon-traversal 
fix: wildcard CORS + credentials spec violation, path-traversal guard on /icons proxy
 2026-07-31 19:27:20 +03:00
Alexander Myasoedov 42615e506a fix(build): 2026-06-23 10:20:10 +03:00
Alexander Myasoedov e6459a551a Merge pull request #321 from DevamShah/config-pluggable-detectors 
feat: config-pluggable refusal classifiers and leak detectors
 2026-06-23 10:12:26 +03:00
Alexander Myasoedov 8e12141df8 Merge pull request #318 from nakshaatraa/docs/fuzzer-module-docstring 
docs: add module-level docstring and document constants in fuzzer.py
 2026-06-15 12:55:19 +03:00
Alexander Myasoedov b90b80a0af Merge pull request #317 from nakshaatraa/fix/image-generator-matplotlib-warnings 
fix: set matplotlib Agg backend and sanitize prompt whitespace
 2026-06-15 12:54:41 +03:00
Alexander Myasoedov b827a0b186 Merge pull request #316 from jasoncobra3/chore/delete-agno-dead-code-v2 
chore: delete Agno dead code (Phase 1)
 2026-06-15 12:52:19 +03:00
Alexander Myasoedov 21f7517ef9 Merge pull request #314 from JackSpiece/chore/remove-mcp 
chore: delete MCP server and client
 2026-06-11 17:46:04 +03:00
Alexander Myasoedov 438f30bfb2 Merge pull request #313 from JackSpiece/chore/remove-agno 
chore: remove leftover Agno artifacts
 2026-06-10 23:55:45 +03:00
Alexander Myasoedov 13b03b958f Merge pull request #310 from zhanz5/fix/cost-calculation-model-aware 
fix: make cost calculation model-aware instead of hardcoded to deepseek-chat
 2026-06-05 10:12:41 +03:00
Alexander Myasoedov 6ae9ea8bfe fix(pc): 2026-06-04 18:32:42 +03:00
Alexander Myasoedov 40a8284656 feat(clean readme): 2026-06-04 18:29:25 +03:00
Alexander Myasoedov ead8f85836 feat(feat(refusal): detect Docker/K8s sandbox escape probes (#280)): 2026-06-04 18:28:12 +03:00
Alexander Myasoedov 6dcda7c931 fix(fix(security): bind server to 127.0.0.1 instead of 0.0.0.0 by default): 2026-06-04 17:53:35 +03:00
Alexander Myasoedov 7b8d238254 Merge pull request #305 from zhanz5/fix/remove-duplicate-probedataset-msj 
fix: remove duplicate ProbeDataset class from msj_data.py
 2026-06-04 17:47:17 +03:00
Alexander Myasoedov 3b26e57b9e fix(pc): 2026-06-03 15:13:19 +03:00
Alexander Myasoedov 5ce4ed5d91 Merge pull request #301 from JackSpiece/fix/tailwind-v4-static-classes 
fix: migrate static UI to Tailwind v4
 2026-06-03 15:10:49 +03:00
Alexander Myasoedov 816c8c6bc7 fix(make litellm optional import): 2026-06-03 15:08:23 +03:00
Alexander Myasoedov a193ef9c2c fix(pc): 2026-06-03 15:05:59 +03:00
Alexander Myasoedov 67cedfb116 Merge pull request #299 from RheagalFire/feat/add-litellm-provider 
feat: add LiteLLM as provider for 100+ LLM backends
 2026-06-03 15:04:18 +03:00
Alexander Myasoedov 50266554fe Merge pull request #297 from ykd007/feat/mcp-claude-docs-193 
document Claude MCP usage in README
 2026-06-03 15:02:59 +03:00
Alexander Myasoedov 1fa66bd292 Merge pull request #300 from JackSpiece/fix/mcp-client-usage-examples 
docs: add MCP client usage examples
 2026-06-03 15:01:02 +03:00
Alexander Myasoedov 1bfb7dcc20 fix(use_agg_backend): 2026-06-03 14:59:43 +03:00
Alexander Myasoedov c0322d1130 Merge pull request #303 from Carlos-Projects/feat/agno-mcp-integration 
feat: add MCP+Agno integration docs and report chart tests
 2026-06-03 14:55:58 +03:00
Alexander Myasoedov a47543e5e0 Merge pull request #304 from zhanz5/fix/scan-csv-route-implementation 
fix: implement scan-csv route to actually use uploaded CSV data
 2026-06-03 14:54:28 +03:00
Alexander Myasoedov 47ee4f09a6 fix(security alerts): 2026-06-03 14:52:45 +03:00
Alexander Myasoedov bcc12a9443 fix(cost fn): 2026-06-03 14:52:29 +03:00
Alexander Myasoedov d2bbad32b4 fix(fmt): 2026-05-27 22:05:37 +03:00
Alexander Myasoedov 40e59860c0 Merge pull request #302 from Carlos-Projects/main 
fix: strip Content-Length and make hardcoded values configurable (closes #139, #167)
 2026-05-27 22:00:20 +03:00
Alexander Myasoedov 2aabcef414 Merge pull request #291 from Dawn-Fighter/feat/pii-leak-detector 
feat: add PII leak detector
 2026-05-14 20:25:09 +03:00
Alexander Myasoedov 2896974005 fix(pc): 2026-05-14 19:23:22 +03:00
Alexander Myasoedov e38365c904 Merge pull request #290 from ykd007/feat/google-sheets-dataset-support 
feat(datasets): support Google Sheets URLs in dataset loader
 2026-05-14 19:20:49 +03:00
Alexander Myasoedov 669169bd11 feat(vercel): 2026-05-14 19:19:44 +03:00
Alexander Myasoedov cb64a3b70c fix(script): 2026-05-14 19:16:51 +03:00
Alexander Myasoedov e2bf837e00 fix(docs): 2026-05-14 19:15:12 +03:00
Alexander Myasoedov 0fba1ccadf feat(docs/vercel.sh): 2026-05-14 19:14:09 +03:00
Alexander Myasoedov d5cd85f8cb fix(ci): 2026-05-14 19:08:40 +03:00
Alexander Myasoedov 695eac4144 fix(build): 2026-05-14 19:04:16 +03:00
Alexander Myasoedov 50785b9850 fix(pc): 2026-05-14 18:59:37 +03:00
Alexander Myasoedov a0b2b9ec70 feat(py upgrade): 2026-05-14 18:56:24 +03:00
Alexander Myasoedov 46ec775266 feat(deprecate ui): 2026-05-14 18:40:27 +03:00
Alexander Myasoedov 5b90eb032a Merge pull request #277 from niveshdandyan/add-full-event-log 
feat: add full event log export for all scan events
 2026-02-03 18:54:09 +02:00
Alexander Myasoedov 796bd33432 Merge pull request #276 from msoedov/feat/research-enhancements 
Feat/research enhancements
 2026-01-28 21:09:00 +02:00
Alexander Myasoedov bc7fdd7cfa fix(pc): 2026-01-28 21:04:29 +02:00
Alexander Myasoedov 8d42a84a9d fix(cleanup): 2026-01-28 21:00:08 +02:00
Alexander Myasoedov 49b2243258 docs: Update PRD and progress for US-005 completion 2026-01-28 18:52:42 +02:00
Alexander Myasoedov b38a27d78c feat: US-005 - Enhanced Refusal Detection with Hybrid Approach 
Implement hybrid refusal classifier combining multiple detection methods:
- Add confidence scoring to refusal detection (HybridResult)
- Implement weighted voting with configurable thresholds
- Support require_unanimous mode for strict classification
- Add factory function create_hybrid_classifier for common setup
- Include 32 unit tests with table-driven test patterns
 2026-01-28 18:52:20 +02:00
Alexander Myasoedov d5e2746567 docs: Update PRD and progress for US-004 completion 2026-01-28 18:35:07 +02:00
Alexander Myasoedov 41567925aa feat: US-004 - Unified LLM Provider Abstraction 
Create unified provider abstraction layer for direct LLM integrations beyond
HTTP specs, inspired by FuzzyAI's comprehensive provider system.

- Add BaseLLMProvider abstract class with standard interface (generate, chat,
  sync_generate, sync_chat methods)
- Implement OpenAIProvider supporting chat completions API
- Implement AnthropicProvider supporting messages API
- Create provider factory for instantiation by name (create_provider,
  get_provider_class)
- Add 60 unit tests covering all provider implementations
 2026-01-28 18:34:38 +02:00
Alexander Myasoedov 29decc5c4e docs: Update PRD and progress for US-003 completion 2026-01-28 18:29:44 +02:00
Alexander Myasoedov f8e3f6f4a5 feat: US-003 - Composable Fuzzing Chain System 
Implement FuzzNode and FuzzChain classes for multi-step attack chains
with pipe operator syntax, inspired by FuzzyAI architecture.

- FuzzNode: Single LLM call with {var} template substitution
- FuzzChain: Sequential execution passing output as input
- Pipe operator (|) for composing nodes into chains
- LLMProvider protocol for provider abstraction
- 22 unit tests covering composition and execution
 2026-01-28 18:29:22 +02:00
Alexander Myasoedov d5ec249b6c docs: Update PRD and progress for US-002 completion 2026-01-28 18:23:30 +02:00
Alexander Myasoedov ef35c1f82e feat: US-002 - YAML-based Attack Rule System 
Implement a YAML-based rule system for defining attack patterns and success
conditions, inspired by Promptmap's 50+ YAML rule definitions.

Features:
- AttackRule model with name, type, severity, prompt, pass/fail conditions
- RuleLoader for parsing YAML files with validation
- Support for recursive directory loading and filtering by type/severity
- Template variable substitution in prompts
- Dataset integration for converting rules to ProbeDataset format
- YAMLRulesDatasetLoader for loading rules from multiple directories

Tested with 47 unit tests covering models, loader, and dataset integration.
Successfully loads 69 rules from promptmap research directory.
 2026-01-28 18:23:04 +02:00
Alexander Myasoedov 93a85029cb docs: Update PRD and progress for US-001 completion 2026-01-28 18:18:32 +02:00
Alexander Myasoedov 32f103acbc feat: US-001 - Dual-LLM Evaluation for Attack Success Detection 
Add LLM-based refusal classifier inspired by Promptmap's dual-LLM
architecture. The controller LLM evaluates whether an attack succeeded
by analyzing the target's response against pass/fail conditions.

- Create LLMRefusalClassifier plugin integrating with existing system
- Support OpenAI and Anthropic providers with lazy initialization
- Add configurable system prompts and pass/fail conditions
- Include 20 unit tests for comprehensive coverage
 2026-01-28 18:18:09 +02:00
Alexander Myasoedov ce7636fe9e feat(restruct tests): 2025-12-26 22:58:21 +02:00
Alexander Myasoedov 433c999600 feat(clean up obsolete agents): 2025-12-24 08:21:48 +02:00
Alexander Myasoedov 268f9ecf86 Merge pull request #275 from msoedov/poc-concurrency-reporting-unified 
Poc concurrency reporting, general improvements
 2025-12-24 08:20:42 +02:00
Alexander Myasoedov 5238d67846 feat(cleanup): 2025-12-24 08:18:17 +02:00
Alexander Myasoedov a9adb22458 fix(pc): 2025-12-24 08:16:21 +02:00
Alexander Myasoedov 2dc41af98d feat(cleanup): 2025-12-24 08:11:43 +02:00
Alexander Myasoedov 48125bd106 feat(add executor): 2025-12-24 08:10:08 +02:00
Alexander Myasoedov 5285fdd0a0 codex quality run #1 2025-12-10 23:06:40 +02:00
Alexander Myasoedov bf628db5c4 codex quality run #1 2025-12-10 22:53:55 +02:00
Alexander Myasoedov d56b406e1a fix(tests runtime): 2025-12-09 20:00:04 +02:00
Alexander Myasoedov b9dc5de708 feat(add cache dir): 2025-12-09 19:51:47 +02:00
Alexander Myasoedov 9a4fb05491 fix(pc): 2025-11-30 18:50:00 +02:00
Alexander Myasoedov 3e2df49976 fix(pc): 2025-11-30 18:47:15 +02:00
Alexander Myasoedov 14eefb7a67 fix(clean up): 2025-11-30 18:43:37 +02:00
Alexander Myasoedov 7a9c884333 fix(pc): 2025-11-30 18:41:00 +02:00
Alexander Myasoedov a8b5876883 fix(ga): 2025-11-30 18:38:41 +02:00
Alexander Myasoedov fbe9885c0b fix(simplify workflow): 2025-11-30 18:37:23 +02:00
Alexander Myasoedov 583eec1a67 fix(gh): 2025-11-30 18:36:33 +02:00
Alexander Myasoedov f19664f95c fix(pc): 2025-11-30 18:32:58 +02:00
Alexander Myasoedov b3ae0026fb fix(warnings): 2025-11-30 18:30:55 +02:00
Alexander Myasoedov 8ddfec303f feat(poetry update): 2025-11-30 14:21:20 +02:00
Alexander Myasoedov c45778f196 Merge pull request #252 from Davda-James/feat/mcp_client_logging 
logging added for mcp client operations
 2025-08-21 15:00:22 +03:00