Skip to content

AGENTS.md

Latest commit

 

History

History
231 lines (168 loc) · 12.2 KB

AGENTS.md

File metadata and controls

231 lines (168 loc) · 12.2 KB
Raw

Agent Workflow for SecretZero (Monorepo)

This document guides AI coding agents and human contributors when working on SecretZero or using it within any repository.
SecretZero is a schema-driven secrets-as-code tool with multiple integrated components: CLI, REST API (secretzero-api), skills framework, Pydantic models, lockfile engine, provider registry, Terraform modules, and documentation.

Start every session by reading .mex/ROUTER.md for current project state and active patterns.

Core Principles (Apply Across All Sub-Projects)

  • Schema-driven development is mandatory: All changes to configuration, validation, or behavior must flow through the Pydantic models in src/secretzero/models.py (or equivalent). Run task schema:update after any model changes to regenerate Secretfile.schema.json and keep CLI, API, skills, and examples in parity.
  • Absolute rule — SecretZero only, never secrets in context: Agents must not consume secret values in LLM/tool context (no paste in chat, no reading .env/credential files into context, no get --reveal / render / value dumps). Use SecretZero end-to-end for discovery (metadata-only), authoring (placeholders), seeding (agent sync --web, secretzero web), sync, and audit. Canonical wording: skills/secretzero/SKILL.md.
  • Zero-leakage rule (same mandate): Never request, receive, log, store, or allow plaintext secret values into any agent/LLM context, history, logs, or responses.
  • Feature parity: Ensure new features (especially agentic capabilities) are implemented consistently in the CLI (secretzero command), the FastAPI REST API, the skills system, and any related sub-projects (Terraform, docs, .mex patterns).
  • Prefer JSON output (--json for CLI, appropriate API endpoints) for machine-readable results.
  • After model/schema changes, always update:
    • CLI handlers
    • API routers/endpoints
    • Skills documentation
    • Examples (Secretfile.example.yml, tests)
    • .mex/ patterns and ROUTER.md

Unified Agentic Secret-Zero Workflow

The project provides one primary unified entrypoint for secure secret bootstrapping that covers the three agentic vectors while maintaining full compatibility with all other SecretZero workflows (sync, rotate, drift, policy enforcement, audit, etc.).

Primary command (CLI):

secretzero agent sync --json [--web] [--dry-run] [--verbose]

Equivalent API usage (when the REST API is deployed):

  • POST /agent/sync with JSON body { "dry_run", "web", "lockfile?", "sz_agent?" } (same semantics as the CLI); poll GET /agent/sync/web/{session_id} after Vector 2.

The Three Agentic Vectors

Vector 1 – Agent instructs human (CLI-guided / human-seeded)
Agent receives rendered instructions and relays them to the human. Human performs actions and confirms.
Best for environments where the human has direct access to consoles or local tools.

Vector 2 – Secure human input via temporary local web UI
Agent triggers --web (or API equivalent). A localhost-only, one-time form collects values that are fed directly into the sync engine. Values never enter agent context or logs.

Vector 3 – Fully automated
Provider authentication is sufficient (or SZ_AGENT=true is set). The workflow completes end-to-end with no human intervention. Failures are reported clearly in structured output.

Universal Agent Loop (recommended for all three vectors):

  1. Call secretzero agent sync --json (or API equivalent).
  2. Act on structured results (pending_secrets, failed_secrets, status).
  3. For Vector 1: Present templated instructions.
  4. For Vector 2: Trigger web UI and guide user safely.
  5. For Vector 3: Proceed automatically.
  6. Re-run until clean.
  7. Only continue with downstream tasks once secret-zero state is resolved.

Full details now live in focused skills:

  • ./skills/secretzero-agent/SKILL.md for unified agentic/runtime workflows and vector handling.
  • ./skills/secretzero-author/SKILL.md for schema-compliant Secretfile.yml authoring and safe discovery-driven policy authoring.

Entra Agent ID workflow extension

When authoring or syncing kind: entra-agent-blueprint secrets:

  • Prefer secretzero agent sync --json first, then escalate to --web for sponsor/manual approval paths.
  • Never expose Graph-generated client secret plaintext in agent output.
  • Ensure Graph permissions are present before retry loops:
    • AgentIdentityBlueprint.Create
    • AgentIdentityBlueprint.AddRemoveCreds.All
    • AgentIdentityBlueprint.UpdateAuthProperties.All
    • Application.ReadWrite.All
    • Directory.ReadWrite.All

Schema-Driven Development Mandate

  • All configuration, validation rules, and agent behavior are defined in Pydantic models.
  • After any model change → run task schema:update.
  • This automatically keeps:
    • CLI argument parsing and output
    • FastAPI endpoints and OpenAPI spec
    • JSON Schema for IDEs and validation
    • Skills and example files
    • Tests and Terraform modules (where applicable)
  • in sync. Never implement behavior directly in CLI or API without updating the core models first.

Pre-Push / Pre-Merge Checklist (Monorepo-Wide)

Run these from the repository root before any push or merge request:

./scripts/agent.pre-commit.sh --mode full
  • If schema:update, lint:fix, or format modifies files, the script exits with guidance. Commit those changes and re-run at least task test and task security:scan.
  • Verify feature parity: Test changes through both CLI and API (if the API sub-component is active).
  • Update .mex/ROUTER.md and any affected patterns in .mex/patterns/.

End-of-Work Prompt

At the end of any coding work, run a fast local gate:

./scripts/agent.pre-commit.sh --mode fast --quiet

The gate runs task docs:links (lychee hyperlink check on README.md and docs/). For doc-only edits: task docs:links.

Before any push/merge, run the full gate:

./scripts/agent.pre-commit.sh --mode full

Other Workflows (Beyond the Three Agentic Vectors)

SecretZero supports many additional workflows that agents should use when appropriate:

  • Standard sync/rotate/drift/policy checks
  • Provider-specific operations
  • Audit log querying
  • Graph visualization of secret dependencies
  • Terraform integration for infrastructure secrets
  • CI/CD secret injection

Always choose the most appropriate workflow; fall back to the unified agent sync only for secret-zero bootstrapping scenarios.

Merge Requests / Pull Requests

  • Include a clear description noting impacts on schema, CLI, API, skills, and any other sub-projects.
  • Confirm schema-driven changes were validated with task schema:update and full tests.
  • Reference updated documentation (AGENTS.md, CLAUDE.md, SKILL.md, docs/).

This ensures SecretZero remains secure, consistent, and maintainable across its entire monorepo surface.

GitNexus — Code Intelligence

This project is indexed by GitNexus as SecretZero (10500 symbols, 21810 relationships, 300 execution flows). Use the GitNexus MCP tools to understand code, assess impact, and navigate safely.

If any GitNexus tool warns the index is stale, run npx gitnexus analyze in terminal first.

Always Do

  • MUST run impact analysis before editing any symbol. Before modifying a function, class, or method, run gitnexus_impact({target: "symbolName", direction: "upstream"}) and report the blast radius (direct callers, affected processes, risk level) to the user.
  • MUST run gitnexus_detect_changes() before committing to verify your changes only affect expected symbols and execution flows.
  • MUST warn the user if impact analysis returns HIGH or CRITICAL risk before proceeding with edits.
  • When exploring unfamiliar code, use gitnexus_query({query: "concept"}) to find execution flows instead of grepping. It returns process-grouped results ranked by relevance.
  • When you need full context on a specific symbol — callers, callees, which execution flows it participates in — use gitnexus_context({name: "symbolName"}).

When Debugging

  1. gitnexus_query({query: "<error or symptom>"}) — find execution flows related to the issue
  2. gitnexus_context({name: "<suspect function>"}) — see all callers, callees, and process participation
  3. READ gitnexus://repo/SecretZero/process/{processName} — trace the full execution flow step by step
  4. For regressions: gitnexus_detect_changes({scope: "compare", base_ref: "main"}) — see what your branch changed

When Refactoring

  • Renaming: MUST use gitnexus_rename({symbol_name: "old", new_name: "new", dry_run: true}) first. Review the preview — graph edits are safe, text_search edits need manual review. Then run with dry_run: false.
  • Extracting/Splitting: MUST run gitnexus_context({name: "target"}) to see all incoming/outgoing refs, then gitnexus_impact({target: "target", direction: "upstream"}) to find all external callers before moving code.
  • After any refactor: run gitnexus_detect_changes({scope: "all"}) to verify only expected files changed.

Never Do

  • NEVER edit a function, class, or method without first running gitnexus_impact on it.
  • NEVER ignore HIGH or CRITICAL risk warnings from impact analysis.
  • NEVER rename symbols with find-and-replace — use gitnexus_rename which understands the call graph.
  • NEVER commit changes without running gitnexus_detect_changes() to check affected scope.

Tools Quick Reference

Tool When to use Command
query Find code by concept gitnexus_query({query: "auth validation"})
context 360-degree view of one symbol gitnexus_context({name: "validateUser"})
impact Blast radius before editing gitnexus_impact({target: "X", direction: "upstream"})
detect_changes Pre-commit scope check gitnexus_detect_changes({scope: "staged"})
rename Safe multi-file rename gitnexus_rename({symbol_name: "old", new_name: "new", dry_run: true})
cypher Custom graph queries gitnexus_cypher({query: "MATCH ..."})

Impact Risk Levels

Depth Meaning Action
d=1 WILL BREAK — direct callers/importers MUST update these
d=2 LIKELY AFFECTED — indirect deps Should test
d=3 MAY NEED TESTING — transitive Test if critical path

Resources

Resource Use for
gitnexus://repo/SecretZero/context Codebase overview, check index freshness
gitnexus://repo/SecretZero/clusters All functional areas
gitnexus://repo/SecretZero/processes All execution flows
gitnexus://repo/SecretZero/process/{name} Step-by-step execution trace

Self-Check Before Finishing

Before completing any code modification task, verify:

  1. gitnexus_impact was run for all modified symbols
  2. No HIGH/CRITICAL risk warnings were ignored
  3. gitnexus_detect_changes() confirms changes match expected scope
  4. All d=1 (WILL BREAK) dependents were updated

Keeping the Index Fresh

After committing code changes, the GitNexus index becomes stale. Re-run analyze to update it:

npx gitnexus analyze

If the index previously included embeddings, preserve them by adding --embeddings:

npx gitnexus analyze --embeddings

To check whether embeddings exist, inspect .gitnexus/meta.json — the stats.embeddings field shows the count (0 means no embeddings). Running analyze without --embeddings will delete any previously generated embeddings.

Claude Code users: A PostToolUse hook handles this automatically after git commit and git merge.

CLI

Task Read this skill file
Understand architecture / "How does X work?" .claude/skills/gitnexus/gitnexus-exploring/SKILL.md
Blast radius / "What breaks if I change X?" .claude/skills/gitnexus/gitnexus-impact-analysis/SKILL.md
Trace bugs / "Why is X failing?" .claude/skills/gitnexus/gitnexus-debugging/SKILL.md
Rename / extract / split / refactor .claude/skills/gitnexus/gitnexus-refactoring/SKILL.md
Tools, resources, schema reference .claude/skills/gitnexus/gitnexus-guide/SKILL.md
Index, status, clean, wiki CLI commands .claude/skills/gitnexus/gitnexus-cli/SKILL.md