thoughtjack 0.6.0

Adversarial agent security testing tool
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226

name: E2E Nightly

on:
  schedule:
    - cron: "0 3 * * *"
  workflow_dispatch:

permissions: read-all

env:
  CARGO_TERM_COLOR: always

jobs:
  build-tj:
    name: Build ThoughtJack
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          submodules: true
      - uses: dtolnay/rust-toolchain@e814c742d4444ce2f3f6abddea7faf00161ed941 # 1.88
      - uses: Swatinem/rust-cache@23869a5bd66c73db3c0ac40331f3206eb23791dc # v2.9.1
      - name: Build release binary
        run: cargo build --release
      - uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f # v7.0.0
        with:
          name: thoughtjack-binary
          path: target/release/thoughtjack
          retention-days: 3

  e2e-pinned:
    name: "pinned: ${{ matrix.scenario }} × ${{ matrix.framework }}"
    runs-on: ubuntu-latest
    needs: [build-tj]
    continue-on-error: true
    strategy:
      fail-fast: false
      matrix:
        include:
          # Framework tests (3)
          - scenario: mcp-tool-invocation
            framework: langgraph
            self_test: false
          - scenario: mcp-tool-invocation
            framework: crewai
            self_test: false
          - scenario: a2a-task-delegation
            framework: crewai
            self_test: false
          # Self-tests (12)
          - scenario: mcp-client-basic
            framework: self-test
            self_test: true
          - scenario: a2a-client-basic
            framework: self-test
            self_test: true
          - scenario: mcp-resources
            framework: self-test
            self_test: true
          - scenario: mcp-prompts
            framework: self-test
            self_test: true
          - scenario: multi-phase-rug-pull
            framework: self-test
            self_test: true
          - scenario: extractor-cross-actor
            framework: self-test
            self_test: true
          - scenario: cel-indicators
            framework: self-test
            self_test: true
          - scenario: verdict-not-exploited
            framework: self-test
            self_test: true
          - scenario: verdict-all-correlation
            framework: self-test
            self_test: true
          - scenario: a2a-streaming
            framework: self-test
            self_test: true
          # mcp-side-effects disabled: server panics with HTTP transport
          # during logging/sampling interleaving (needs stdio transport)
          - scenario: mcp-instructions
            framework: self-test
            self_test: true
          - scenario: error-graceful-degradation
            framework: self-test
            self_test: true
    steps:
      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2

      - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4
        with:
          name: thoughtjack-binary
          path: target/release
      - name: Make binary executable
        run: chmod +x target/release/thoughtjack

      - uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
        with:
          python-version: "3.12"

      - uses: actions/setup-node@53b83947a5a98c8d113130e565377fae1a50d02f # v6.3.0
        if: ${{ !matrix.self_test }}
        with:
          node-version: "22"

      - name: Install mock-llm
        if: ${{ !matrix.self_test }}
        run: npm install -g @dwmkerr/mock-llm@0.1.28

      - name: Start mock-llm
        if: ${{ !matrix.self_test }}
        run: |
          mock-llm &
          echo "MOCK_LLM_PID=$!" >> "$GITHUB_ENV"
          sleep 2

      - name: Install agent dependencies (pinned)
        if: ${{ !matrix.self_test }}
        run: pip install -e tests/e2e/reference-agents/${{ matrix.framework }}/

      - name: Install orchestrator dependencies
        run: pip install "pyyaml==6.0.3"

      - name: Run e2e scenario
        run: |
          args=(
            --scenario "tests/e2e/fixtures/${{ matrix.scenario }}"
            --tj-binary ./target/release/thoughtjack
            --base-port 19000
            --timeout 60
            --output-dir tests/e2e/results
          )
          if [ "${{ matrix.self_test }}" = "true" ]; then
            args+=(--self-test)
          else
            args+=(
              --framework "${{ matrix.framework }}"
              --mock-llm-url http://localhost:6556
            )
          fi
          python tests/e2e/run_conformance.py "${args[@]}"

      - name: Upload results
        if: always()
        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f # v7.0.0
        with:
          name: "e2e-pinned-${{ matrix.scenario }}-${{ matrix.framework }}"
          path: tests/e2e/results/
          retention-days: 14

      - name: Stop mock-llm
        if: always() && !matrix.self_test
        run: kill "$MOCK_LLM_PID" 2>/dev/null || true

  e2e-upgrade:
    name: "upgrade: ${{ matrix.scenario }} × ${{ matrix.framework }}"
    runs-on: ubuntu-latest
    needs: [build-tj]
    continue-on-error: true
    strategy:
      fail-fast: false
      matrix:
        include:
          - scenario: mcp-tool-invocation
            framework: langgraph
          - scenario: mcp-tool-invocation
            framework: crewai
          - scenario: a2a-task-delegation
            framework: crewai
    steps:
      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2

      - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4
        with:
          name: thoughtjack-binary
          path: target/release
      - name: Make binary executable
        run: chmod +x target/release/thoughtjack

      - uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
        with:
          python-version: "3.12"

      - uses: actions/setup-node@53b83947a5a98c8d113130e565377fae1a50d02f # v6.3.0
        with:
          node-version: "22"

      - name: Install mock-llm
        run: npm install -g @dwmkerr/mock-llm@0.1.28

      - name: Start mock-llm
        run: |
          mock-llm &
          echo "MOCK_LLM_PID=$!" >> "$GITHUB_ENV"
          sleep 2

      - name: Install agent dependencies (--upgrade)
        run: pip install --upgrade -e tests/e2e/reference-agents/${{ matrix.framework }}/

      - name: Install orchestrator dependencies
        run: pip install "pyyaml==6.0.3"

      - name: Run e2e scenario
        run: |
          python tests/e2e/run_conformance.py \
            --scenario "tests/e2e/fixtures/${{ matrix.scenario }}" \
            --framework "${{ matrix.framework }}" \
            --tj-binary ./target/release/thoughtjack \
            --mock-llm-url http://localhost:6556 \
            --base-port 19000 \
            --timeout 60 \
            --output-dir tests/e2e/results

      - name: Upload results
        if: always()
        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f # v7.0.0
        with:
          name: "e2e-upgrade-${{ matrix.scenario }}-${{ matrix.framework }}"
          path: tests/e2e/results/
          retention-days: 14

      - name: Stop mock-llm
        if: always()
        run: kill "$MOCK_LLM_PID" 2>/dev/null || true