aip-protocol 0.1.0__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (18) hide show
  1. aip_protocol-0.1.0/.dockerignore +23 -0
  2. aip_protocol-0.1.0/.gitignore +43 -0
  3. aip_protocol-0.1.0/LICENSE +21 -0
  4. aip_protocol-0.1.0/PKG-INFO +283 -0
  5. aip_protocol-0.1.0/README.md +259 -0
  6. aip_protocol-0.1.0/RFC-001.md +225 -0
  7. aip_protocol-0.1.0/aip_protocol/__init__.py +49 -0
  8. aip_protocol-0.1.0/aip_protocol/cli.py +207 -0
  9. aip_protocol-0.1.0/aip_protocol/crypto.py +132 -0
  10. aip_protocol-0.1.0/aip_protocol/envelope.py +165 -0
  11. aip_protocol-0.1.0/aip_protocol/errors.py +93 -0
  12. aip_protocol-0.1.0/aip_protocol/models.py +205 -0
  13. aip_protocol-0.1.0/aip_protocol/passport.py +233 -0
  14. aip_protocol-0.1.0/aip_protocol/revocation.py +209 -0
  15. aip_protocol-0.1.0/aip_protocol/trust.py +133 -0
  16. aip_protocol-0.1.0/aip_protocol/verification.py +504 -0
  17. aip_protocol-0.1.0/pyproject.toml +41 -0
  18. aip_protocol-0.1.0/tests/test_aip.py +996 -0
aip_protocol-0.1.0/.dockerignore ADDED
@@ -0,0 +1,23 @@
1
+ __pycache__/
2
+ *.pyc
3
+ *.pyo
4
+ *.pyd
5
+ .Python
6
+ env/
7
+ venv/
8
+ .env
9
+ .venv
10
+ pip-log.txt
11
+ pip-delete-this-directory.txt
12
+ .tox/
13
+ .coverage
14
+ .coverage.*
15
+ .cache
16
+ nosetests.xml
17
+ coverage.xml
18
+ *.cover
19
+ *.log
20
+ .git
21
+ .mypy_cache
22
+ .pytest_cache
23
+ .hypotheses
aip_protocol-0.1.0/.gitignore ADDED
@@ -0,0 +1,43 @@
1
+ # ── Commercial Product (PRIVATE — never push) ──
2
+ kya_api/
3
+ dashboard/
4
+ *.pem
5
+ Dockerfile
6
+ docker-compose.yml
7
+ Caddyfile
8
+ start.sh
9
+ setup.sh
10
+ data/
11
+
12
+ # ── Logs & Runtime ──
13
+ *.log
14
+ startup.log
15
+ backend.log
16
+ frontend.log
17
+ nohup.out
18
+
19
+ # ── Python ──
20
+ __pycache__/
21
+ *.pyc
22
+ *.pyo
23
+ *.egg-info/
24
+ dist/
25
+ build/
26
+ *.whl
27
+ .eggs/
28
+
29
+ # ── Environment ──
30
+ .env
31
+ .venv/
32
+ venv/
33
+ env/
34
+
35
+ # ── IDE ──
36
+ .vscode/
37
+ .idea/
38
+ *.swp
39
+ *.swo
40
+
41
+ # ── OS ──
42
+ .DS_Store
43
+ Thumbs.db
aip_protocol-0.1.0/LICENSE ADDED
@@ -0,0 +1,21 @@
1
+ MIT License
2
+
3
+ Copyright (c) 2026 KYA Labs
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.
aip_protocol-0.1.0/PKG-INFO ADDED
@@ -0,0 +1,283 @@
1
+ Metadata-Version: 2.4
2
+ Name: aip-protocol
3
+ Version: 0.1.0
4
+ Summary: Agent Intent Protocol — Proof of Intent for the Agentic Web
5
+ Author-email: Aniket Giri <aniket@kyalabs.com>
6
+ License-Expression: MIT
7
+ License-File: LICENSE
8
+ Keywords: agents,ai,identity,protocol,trust,verification
9
+ Classifier: Development Status :: 3 - Alpha
10
+ Classifier: Intended Audience :: Developers
11
+ Classifier: License :: OSI Approved :: MIT License
12
+ Classifier: Programming Language :: Python :: 3
13
+ Classifier: Topic :: Security :: Cryptography
14
+ Classifier: Topic :: Software Development :: Libraries :: Python Modules
15
+ Requires-Python: >=3.10
16
+ Requires-Dist: click>=8.0
17
+ Requires-Dist: cryptography>=42.0
18
+ Requires-Dist: pydantic>=2.0
19
+ Requires-Dist: rich>=13.0
20
+ Provides-Extra: dev
21
+ Requires-Dist: pytest-asyncio>=0.23; extra == 'dev'
22
+ Requires-Dist: pytest>=8.0; extra == 'dev'
23
+ Description-Content-Type: text/markdown
24
+
25
+ <p align="center">
26
+ <img src="https://img.shields.io/badge/Protocol-AIP--1-0B0D10?style=for-the-badge&labelColor=0B0D10&color=ABDBE3" alt="AIP-1" />
27
+ <img src="https://img.shields.io/badge/Crypto-Ed25519-0B0D10?style=for-the-badge&labelColor=0B0D10&color=34D399" alt="Ed25519" />
28
+ <img src="https://img.shields.io/pypi/v/aip-protocol?style=for-the-badge&labelColor=0B0D10&color=A78BFA&label=PyPI" alt="PyPI" />
29
+ <img src="https://img.shields.io/badge/Tests-63%20passing-0B0D10?style=for-the-badge&labelColor=0B0D10&color=34D399" alt="Tests" />
30
+ <img src="https://img.shields.io/badge/License-MIT-0B0D10?style=for-the-badge&labelColor=0B0D10&color=94A3B8" alt="License" />
31
+ </p>
32
+
33
+ <h1 align="center">AIP — Agent Intent Protocol</h1>
34
+
35
+ <p align="center">
36
+ <strong>The HTTPS for AI Agents.</strong><br/>
37
+ Cryptographic identity, intent verification, and boundary enforcement for autonomous agents.
38
+ </p>
39
+
40
+ <p align="center">
41
+ <a href="https://aip.synthexai.tech/docs">Documentation</a> ·
42
+ <a href="https://aip.synthexai.tech">Live Dashboard</a> ·
43
+ <a href="https://pypi.org/project/aip-protocol/">PyPI</a>
44
+ </p>
45
+
46
+ ---
47
+
48
+ ## The Problem
49
+
50
+ Every AI framework lets agents **do things**. None of them verify **what agents are allowed to do**.
51
+
52
+ A LangChain agent can drain a bank account. An AutoGPT agent can email your customers. A CrewAI agent can delete production data. There is no standard way to verify an agent's identity, enforce its boundaries, or revoke it in real-time.
53
+
54
+ **AIP fixes this.**
55
+
56
+ ## What is AIP?
57
+
58
+ AIP-1 is a trustless, cross-platform protocol for verifying the **identity**, **intent**, and **authorization boundaries** of autonomous AI agents before they act.
59
+
60
+ Think of it as **OAuth + TLS, purpose-built for the agentic web**.
61
+
62
+ ```
63
+ Agent wants to act → Creates signed Intent Envelope → Verifier checks 8-step pipeline → Allow or Deny
64
+ ```
65
+
66
+ ### Core Capabilities
67
+
68
+ | Capability | What it does |
69
+ |---|---|
70
+ | **Cryptographic Identity** | Ed25519 keypair per agent, DID-based addressing (`did:web:`) |
71
+ | **Boundary Enforcement** | Action allowlists, deny lists, monetary limits, geo restrictions |
72
+ | **Tiered Verification** | Sub-millisecond for low-risk, full crypto for high-value intents |
73
+ | **Kill Switch** | Revoke or suspend any agent globally with zero propagation delay |
74
+ | **Trust Scores** | Bayesian reputation model — trust is earned over successful verifications |
75
+ | **Intent Drift Detection** | Semantic classifier flags actions outside an agent's declared scope |
76
+ | **Structured Error Codes** | 22 machine-readable `AIP-Exxx` codes across 5 categories for audit trails |
77
+
78
+ ## Install
79
+
80
+ ```bash
81
+ pip install aip-protocol
82
+ ```
83
+
84
+ ## Quick Start
85
+
86
+ ```python
87
+ from aip_protocol import AgentPassport, create_envelope, sign_envelope, verify_intent
88
+ from aip_protocol.revocation import RevocationStore
89
+
90
+ # 1 — Create an agent passport (identity + keys + boundaries)
91
+ passport = AgentPassport.create(
92
+ domain="yourco.com",
93
+ agent_name="procurement-bot",
94
+ allowed_actions=["read_invoice", "transfer_funds"],
95
+ monetary_limit_per_txn=50.0,
96
+ )
97
+
98
+ print(passport.agent_id)
99
+ # → "did:web:yourco.com:agents:procurement-bot"
100
+
101
+ # 2 — Agent wants to act: create and sign an intent envelope
102
+ envelope = create_envelope(
103
+ passport,
104
+ action="transfer_funds",
105
+ target="did:web:vendor.com",
106
+ parameters={"amount": 45.00, "currency": "USD"},
107
+ )
108
+ signed = sign_envelope(envelope, passport.private_key)
109
+
110
+ # 3 — Verifier checks the intent through the 8-step pipeline
111
+ store = RevocationStore()
112
+ result = verify_intent(signed, passport.public_key, revocation_store=store)
113
+
114
+ if result.passed:
115
+ print(f"✓ Verified — tier: {result.tier_used.value}, trust: {result.trust_score}")
116
+ else:
117
+ for error in result.errors:
118
+ print(f"✗ {error.value}: {error.name}")
119
+ # e.g. "✗ AIP-E202: MONETARY_LIMIT"
120
+ ```
121
+
122
+ ## Verification Pipeline
123
+
124
+ Every intent passes through an 8-step verification pipeline. The protocol auto-selects the verification tier based on risk:
125
+
126
+ ```
127
+ ┌────────────────────────────────────────────────────────┐
128
+ │ Intent Envelope │
129
+ │ ┌───────────┐ ┌───────────┐ ┌────────────────────┐ │
130
+ │ │ Agent ID │ │ Intent │ │ Boundaries │ │
131
+ │ │ (DID) │ │ (Action) │ │ (The Cage) │ │
132
+ │ └─────┬──────┘ └─────┬─────┘ └──────────┬─────────┘ │
133
+ │ └───────────────┼────────────────────┘ │
134
+ │ ┌─────▼─────┐ │
135
+ │ │ Proof │ ← Ed25519 signature │
136
+ │ └───────────┘ │
137
+ └────────────────────────────────────────────────────────┘
138
+ │
139
+ ▼
140
+ ┌────────────────────────────────────────────────────────┐
141
+ │ Verification Pipeline │
142
+ │ │
143
+ │ ① Version Check ⑤ Attestation Verify │
144
+ │ ② Schema Validation ⑥ Revocation Check │
145
+ │ ③ Expiry Check ⑦ Trust Score Evaluation │
146
+ │ ④ Boundary Check ⑧ Final Verdict │
147
+ │ └─ Actions │
148
+ │ └─ Monetary limits │
149
+ │ └─ Geo restrictions │
150
+ │ └─ Intent drift │
151
+ └────────────────────────────────────────────────────────┘
152
+ ```
153
+
154
+ ### Tiered Verification
155
+
156
+ Not every intent needs full cryptographic verification. AIP auto-selects the tier:
157
+
158
+ | Tier | Use Case | Latency | What Runs |
159
+ |---|---|---|---|
160
+ | **Tier 0** | Low-risk, cached, in-session repeats | **<1ms** | HMAC + boundary proof |
161
+ | **Tier 1** | Normal operations | **~5ms** | Ed25519 + boundary + revocation |
162
+ | **Tier 2** | High-value, cross-org, first contact | **~50–100ms** | Full 8-step pipeline |
163
+
164
+ ## Error Taxonomy
165
+
166
+ Every failure returns a machine-readable `AIP-Exxx` code — not a generic 400. Your logs, dashboards, and audit trails show *exactly* what went wrong.
167
+
168
+ | Range | Category | Examples |
169
+ |---|---|---|
170
+ | `AIP-E1xx` | **Envelope Errors** | `E100` Invalid Signature · `E101` Expired · `E102` Replay Detected |
171
+ | `AIP-E2xx` | **Boundary Violations** | `E200` Action Not Allowed · `E202` Monetary Limit · `E204` Geo Restricted |
172
+ | `AIP-E3xx` | **Attestation Failures** | `E300` Model Hash Mismatch · `E303` Intent Drift |
173
+ | `AIP-E4xx` | **Trust Failures** | `E400` Agent Revoked · `E403` Delegation Invalid · `E404` Trust Too Low |
174
+ | `AIP-E5xx` | **Protocol Errors** | `E500` Mesh Unavailable · `E502` Handshake Timeout |
175
+
176
+ Full reference → [aip.synthexai.tech/docs#errors](https://aip.synthexai.tech/docs#errors)
177
+
178
+ ## CLI
179
+
180
+ ```bash
181
+ # Create a passport
182
+ aip create-passport --domain yourco.com --name my-agent \
183
+ -a read_data -a transfer_funds -m 100
184
+
185
+ # Sign an intent
186
+ aip sign-intent --passport ./agent_passport \
187
+ --action transfer_funds --amount 45 -o intent.json
188
+
189
+ # Verify an intent
190
+ aip verify --envelope intent.json \
191
+ --public-key ./agent_passport/public.pem
192
+
193
+ # Revoke an agent instantly
194
+ aip revoke "did:web:yourco.com:agents:my-agent" --reason "compromised"
195
+ ```
196
+
197
+ ## Hosted API
198
+
199
+ Self-host the verification engine, or use the managed cloud at [aip.synthexai.tech](https://aip.synthexai.tech):
200
+
201
+ ```bash
202
+ curl -X POST https://aip.synthexai.tech/api/verify \
203
+ -H "X-API-Key: kya_YOUR_KEY" \
204
+ -H "Content-Type: application/json" \
205
+ -d '{
206
+ "agent_id": "did:web:yourco.com:agents:my-bot",
207
+ "action": "transfer_funds",
208
+ "target": "did:web:vendor.com",
209
+ "parameters": {"amount": 45.00}
210
+ }'
211
+ ```
212
+
213
+ ```json
214
+ {
215
+ "verified": true,
216
+ "tier": "tier_1",
217
+ "signature_valid": true,
218
+ "within_boundaries": true,
219
+ "trust_score": 0.847,
220
+ "latency_ms": 3.21,
221
+ "errors": []
222
+ }
223
+ ```
224
+
225
+ Full API docs → [aip.synthexai.tech/docs](https://aip.synthexai.tech/docs)
226
+
227
+ ## Framework Compatibility
228
+
229
+ AIP is framework-agnostic. It works with any agent runtime:
230
+
231
+ | Framework | Status | Notes |
232
+ |---|---|---|
233
+ | LangChain / LangGraph | ✅ Supported | Framework attestation via `framework_id` |
234
+ | AutoGPT | ✅ Supported | Boundary enforcement on plugin calls |
235
+ | CrewAI | ✅ Supported | Per-agent passport, shared trust mesh |
236
+ | Custom Python agents | ✅ Supported | Any agent can create a passport |
237
+ | OpenAI Assistants | 🔜 Planned | Function-call boundary mapping |
238
+
239
+ ## Development
240
+
241
+ ```bash
242
+ git clone https://github.com/theaniketgiri/aip.git
243
+ cd aip
244
+ pip install -e ".[dev]"
245
+ pytest tests/ -v
246
+ # 63 tests, all passing
247
+ ```
248
+
249
+ ### Project Structure
250
+
251
+ ```
252
+ aip_protocol/
253
+ ├── passport.py # Agent identity + Ed25519 key management
254
+ ├── envelope.py # Intent envelope creation + signing
255
+ ├── verification.py # 8-step verification pipeline + intent classifier
256
+ ├── crypto.py # Ed25519 + HMAC cryptographic layer
257
+ ├── errors.py # AIP-Exxx error taxonomy (22 structured codes)
258
+ ├── revocation.py # Real-time revocation store with rehydration
259
+ ├── trust.py # Bayesian trust score engine
260
+ ├── models.py # Protocol data models (Pydantic)
261
+ └── cli.py # Command-line interface
262
+ ```
263
+
264
+ ## Design Partners
265
+
266
+ We're onboarding **early design partners** building multi-agent systems. Partners get:
267
+
268
+ - Enterprise-tier API access (free during beta)
269
+ - Direct engineering support
270
+ - Protocol roadmap influence
271
+
272
+ → **[Apply at aip.synthexai.tech](https://aip.synthexai.tech)**
273
+
274
+ ## License
275
+
276
+ MIT — see [LICENSE](LICENSE) for details.
277
+
278
+ ---
279
+
280
+ <p align="center">
281
+ <sub><strong>KYA Labs</strong> — Know Your Agent before it acts.</sub><br/>
282
+ <sub><a href="https://aip.synthexai.tech">Website</a> · <a href="https://aip.synthexai.tech/docs">API Docs</a> · <a href="https://pypi.org/project/aip-protocol/">PyPI</a></sub>
283
+ </p>
aip_protocol-0.1.0/README.md ADDED
@@ -0,0 +1,259 @@
1
+ <p align="center">
2
+ <img src="https://img.shields.io/badge/Protocol-AIP--1-0B0D10?style=for-the-badge&labelColor=0B0D10&color=ABDBE3" alt="AIP-1" />
3
+ <img src="https://img.shields.io/badge/Crypto-Ed25519-0B0D10?style=for-the-badge&labelColor=0B0D10&color=34D399" alt="Ed25519" />
4
+ <img src="https://img.shields.io/pypi/v/aip-protocol?style=for-the-badge&labelColor=0B0D10&color=A78BFA&label=PyPI" alt="PyPI" />
5
+ <img src="https://img.shields.io/badge/Tests-63%20passing-0B0D10?style=for-the-badge&labelColor=0B0D10&color=34D399" alt="Tests" />
6
+ <img src="https://img.shields.io/badge/License-MIT-0B0D10?style=for-the-badge&labelColor=0B0D10&color=94A3B8" alt="License" />
7
+ </p>
8
+
9
+ <h1 align="center">AIP — Agent Intent Protocol</h1>
10
+
11
+ <p align="center">
12
+ <strong>The HTTPS for AI Agents.</strong><br/>
13
+ Cryptographic identity, intent verification, and boundary enforcement for autonomous agents.
14
+ </p>
15
+
16
+ <p align="center">
17
+ <a href="https://aip.synthexai.tech/docs">Documentation</a> ·
18
+ <a href="https://aip.synthexai.tech">Live Dashboard</a> ·
19
+ <a href="https://pypi.org/project/aip-protocol/">PyPI</a>
20
+ </p>
21
+
22
+ ---
23
+
24
+ ## The Problem
25
+
26
+ Every AI framework lets agents **do things**. None of them verify **what agents are allowed to do**.
27
+
28
+ A LangChain agent can drain a bank account. An AutoGPT agent can email your customers. A CrewAI agent can delete production data. There is no standard way to verify an agent's identity, enforce its boundaries, or revoke it in real-time.
29
+
30
+ **AIP fixes this.**
31
+
32
+ ## What is AIP?
33
+
34
+ AIP-1 is a trustless, cross-platform protocol for verifying the **identity**, **intent**, and **authorization boundaries** of autonomous AI agents before they act.
35
+
36
+ Think of it as **OAuth + TLS, purpose-built for the agentic web**.
37
+
38
+ ```
39
+ Agent wants to act → Creates signed Intent Envelope → Verifier checks 8-step pipeline → Allow or Deny
40
+ ```
41
+
42
+ ### Core Capabilities
43
+
44
+ | Capability | What it does |
45
+ |---|---|
46
+ | **Cryptographic Identity** | Ed25519 keypair per agent, DID-based addressing (`did:web:`) |
47
+ | **Boundary Enforcement** | Action allowlists, deny lists, monetary limits, geo restrictions |
48
+ | **Tiered Verification** | Sub-millisecond for low-risk, full crypto for high-value intents |
49
+ | **Kill Switch** | Revoke or suspend any agent globally with zero propagation delay |
50
+ | **Trust Scores** | Bayesian reputation model — trust is earned over successful verifications |
51
+ | **Intent Drift Detection** | Semantic classifier flags actions outside an agent's declared scope |
52
+ | **Structured Error Codes** | 22 machine-readable `AIP-Exxx` codes across 5 categories for audit trails |
53
+
54
+ ## Install
55
+
56
+ ```bash
57
+ pip install aip-protocol
58
+ ```
59
+
60
+ ## Quick Start
61
+
62
+ ```python
63
+ from aip_protocol import AgentPassport, create_envelope, sign_envelope, verify_intent
64
+ from aip_protocol.revocation import RevocationStore
65
+
66
+ # 1 — Create an agent passport (identity + keys + boundaries)
67
+ passport = AgentPassport.create(
68
+ domain="yourco.com",
69
+ agent_name="procurement-bot",
70
+ allowed_actions=["read_invoice", "transfer_funds"],
71
+ monetary_limit_per_txn=50.0,
72
+ )
73
+
74
+ print(passport.agent_id)
75
+ # → "did:web:yourco.com:agents:procurement-bot"
76
+
77
+ # 2 — Agent wants to act: create and sign an intent envelope
78
+ envelope = create_envelope(
79
+ passport,
80
+ action="transfer_funds",
81
+ target="did:web:vendor.com",
82
+ parameters={"amount": 45.00, "currency": "USD"},
83
+ )
84
+ signed = sign_envelope(envelope, passport.private_key)
85
+
86
+ # 3 — Verifier checks the intent through the 8-step pipeline
87
+ store = RevocationStore()
88
+ result = verify_intent(signed, passport.public_key, revocation_store=store)
89
+
90
+ if result.passed:
91
+ print(f"✓ Verified — tier: {result.tier_used.value}, trust: {result.trust_score}")
92
+ else:
93
+ for error in result.errors:
94
+ print(f"✗ {error.value}: {error.name}")
95
+ # e.g. "✗ AIP-E202: MONETARY_LIMIT"
96
+ ```
97
+
98
+ ## Verification Pipeline
99
+
100
+ Every intent passes through an 8-step verification pipeline. The protocol auto-selects the verification tier based on risk:
101
+
102
+ ```
103
+ ┌────────────────────────────────────────────────────────┐
104
+ │ Intent Envelope │
105
+ │ ┌───────────┐ ┌───────────┐ ┌────────────────────┐ │
106
+ │ │ Agent ID │ │ Intent │ │ Boundaries │ │
107
+ │ │ (DID) │ │ (Action) │ │ (The Cage) │ │
108
+ │ └─────┬──────┘ └─────┬─────┘ └──────────┬─────────┘ │
109
+ │ └───────────────┼────────────────────┘ │
110
+ │ ┌─────▼─────┐ │
111
+ │ │ Proof │ ← Ed25519 signature │
112
+ │ └───────────┘ │
113
+ └────────────────────────────────────────────────────────┘
114
+ │
115
+ ▼
116
+ ┌────────────────────────────────────────────────────────┐
117
+ │ Verification Pipeline │
118
+ │ │
119
+ │ ① Version Check ⑤ Attestation Verify │
120
+ │ ② Schema Validation ⑥ Revocation Check │
121
+ │ ③ Expiry Check ⑦ Trust Score Evaluation │
122
+ │ ④ Boundary Check ⑧ Final Verdict │
123
+ │ └─ Actions │
124
+ │ └─ Monetary limits │
125
+ │ └─ Geo restrictions │
126
+ │ └─ Intent drift │
127
+ └────────────────────────────────────────────────────────┘
128
+ ```
129
+
130
+ ### Tiered Verification
131
+
132
+ Not every intent needs full cryptographic verification. AIP auto-selects the tier:
133
+
134
+ | Tier | Use Case | Latency | What Runs |
135
+ |---|---|---|---|
136
+ | **Tier 0** | Low-risk, cached, in-session repeats | **<1ms** | HMAC + boundary proof |
137
+ | **Tier 1** | Normal operations | **~5ms** | Ed25519 + boundary + revocation |
138
+ | **Tier 2** | High-value, cross-org, first contact | **~50–100ms** | Full 8-step pipeline |
139
+
140
+ ## Error Taxonomy
141
+
142
+ Every failure returns a machine-readable `AIP-Exxx` code — not a generic 400. Your logs, dashboards, and audit trails show *exactly* what went wrong.
143
+
144
+ | Range | Category | Examples |
145
+ |---|---|---|
146
+ | `AIP-E1xx` | **Envelope Errors** | `E100` Invalid Signature · `E101` Expired · `E102` Replay Detected |
147
+ | `AIP-E2xx` | **Boundary Violations** | `E200` Action Not Allowed · `E202` Monetary Limit · `E204` Geo Restricted |
148
+ | `AIP-E3xx` | **Attestation Failures** | `E300` Model Hash Mismatch · `E303` Intent Drift |
149
+ | `AIP-E4xx` | **Trust Failures** | `E400` Agent Revoked · `E403` Delegation Invalid · `E404` Trust Too Low |
150
+ | `AIP-E5xx` | **Protocol Errors** | `E500` Mesh Unavailable · `E502` Handshake Timeout |
151
+
152
+ Full reference → [aip.synthexai.tech/docs#errors](https://aip.synthexai.tech/docs#errors)
153
+
154
+ ## CLI
155
+
156
+ ```bash
157
+ # Create a passport
158
+ aip create-passport --domain yourco.com --name my-agent \
159
+ -a read_data -a transfer_funds -m 100
160
+
161
+ # Sign an intent
162
+ aip sign-intent --passport ./agent_passport \
163
+ --action transfer_funds --amount 45 -o intent.json
164
+
165
+ # Verify an intent
166
+ aip verify --envelope intent.json \
167
+ --public-key ./agent_passport/public.pem
168
+
169
+ # Revoke an agent instantly
170
+ aip revoke "did:web:yourco.com:agents:my-agent" --reason "compromised"
171
+ ```
172
+
173
+ ## Hosted API
174
+
175
+ Self-host the verification engine, or use the managed cloud at [aip.synthexai.tech](https://aip.synthexai.tech):
176
+
177
+ ```bash
178
+ curl -X POST https://aip.synthexai.tech/api/verify \
179
+ -H "X-API-Key: kya_YOUR_KEY" \
180
+ -H "Content-Type: application/json" \
181
+ -d '{
182
+ "agent_id": "did:web:yourco.com:agents:my-bot",
183
+ "action": "transfer_funds",
184
+ "target": "did:web:vendor.com",
185
+ "parameters": {"amount": 45.00}
186
+ }'
187
+ ```
188
+
189
+ ```json
190
+ {
191
+ "verified": true,
192
+ "tier": "tier_1",
193
+ "signature_valid": true,
194
+ "within_boundaries": true,
195
+ "trust_score": 0.847,
196
+ "latency_ms": 3.21,
197
+ "errors": []
198
+ }
199
+ ```
200
+
201
+ Full API docs → [aip.synthexai.tech/docs](https://aip.synthexai.tech/docs)
202
+
203
+ ## Framework Compatibility
204
+
205
+ AIP is framework-agnostic. It works with any agent runtime:
206
+
207
+ | Framework | Status | Notes |
208
+ |---|---|---|
209
+ | LangChain / LangGraph | ✅ Supported | Framework attestation via `framework_id` |
210
+ | AutoGPT | ✅ Supported | Boundary enforcement on plugin calls |
211
+ | CrewAI | ✅ Supported | Per-agent passport, shared trust mesh |
212
+ | Custom Python agents | ✅ Supported | Any agent can create a passport |
213
+ | OpenAI Assistants | 🔜 Planned | Function-call boundary mapping |
214
+
215
+ ## Development
216
+
217
+ ```bash
218
+ git clone https://github.com/theaniketgiri/aip.git
219
+ cd aip
220
+ pip install -e ".[dev]"
221
+ pytest tests/ -v
222
+ # 63 tests, all passing
223
+ ```
224
+
225
+ ### Project Structure
226
+
227
+ ```
228
+ aip_protocol/
229
+ ├── passport.py # Agent identity + Ed25519 key management
230
+ ├── envelope.py # Intent envelope creation + signing
231
+ ├── verification.py # 8-step verification pipeline + intent classifier
232
+ ├── crypto.py # Ed25519 + HMAC cryptographic layer
233
+ ├── errors.py # AIP-Exxx error taxonomy (22 structured codes)
234
+ ├── revocation.py # Real-time revocation store with rehydration
235
+ ├── trust.py # Bayesian trust score engine
236
+ ├── models.py # Protocol data models (Pydantic)
237
+ └── cli.py # Command-line interface
238
+ ```
239
+
240
+ ## Design Partners
241
+
242
+ We're onboarding **early design partners** building multi-agent systems. Partners get:
243
+
244
+ - Enterprise-tier API access (free during beta)
245
+ - Direct engineering support
246
+ - Protocol roadmap influence
247
+
248
+ → **[Apply at aip.synthexai.tech](https://aip.synthexai.tech)**
249
+
250
+ ## License
251
+
252
+ MIT — see [LICENSE](LICENSE) for details.
253
+
254
+ ---
255
+
256
+ <p align="center">
257
+ <sub><strong>KYA Labs</strong> — Know Your Agent before it acts.</sub><br/>
258
+ <sub><a href="https://aip.synthexai.tech">Website</a> · <a href="https://aip.synthexai.tech/docs">API Docs</a> · <a href="https://pypi.org/project/aip-protocol/">PyPI</a></sub>
259
+ </p>