The AI agent opened a pull request with failing tests. What should I do?

priiism's agent attempts to heal its own failing builds before requesting review. If tests are still failing when the PR opens, leave a comment describing the failure — the agent can iterate. Nothing merges without your approval.

We're concerned about secrets appearing in generated code. Is this handled?

Yes. priiism automatically scans agent output and redacts detected secrets before they appear in pull requests or logs. Per-project secrets are stored separately with encryption at rest.

priiism troubleshooting and common issues

This page covers the most common issues engineering teams encounter when setting up or using priiism, along with how to resolve them and when to escalate.

Repository Connection Issues

Symptom: priiism cannot read repositories after authorization.

Confirm the authorizing user has admin-level access to the GitHub or GitLab organization
Check that the OAuth app or GitHub App was granted access to the correct repositories (not just selected repos that exclude your targets)
Re-authorize if permissions were changed after initial connection

Quality Gates Failing on First Run

Symptom: Typecheck, lint, or build gates fail on AI-generated code.

priiism's agent attempts to self-heal failing builds before requesting review — allow the cycle to complete before intervening
If failures persist, confirm the codebase scan completed fully so priiism has learned your team's standards and linting configuration
Check that your lint and tsconfig files are committed to the repository root where priiism expects them

IDE Plugin Not Surfacing Suggestions

Symptom: Developers installed the plugin but see no AI suggestions.

Verify the plugin is authenticated against the correct priiism organization account
Confirm the repository open in the IDE is connected to priiism
Restart the IDE after plugin installation to ensure proper initialization

Pull Requests Not Opening for GitHub Issues

Symptom: Pointing priiism at a GitHub milestone produces no pull requests.

Confirm the GitHub App has write permission to pull requests in the target repository
Check that issues are in an open state and assigned to the connected project
Review the audit log in the priiism dashboard for any blocked or queued actions awaiting human approval

SSO / SAML Login Failures

Symptom: Developers cannot log in via enterprise SSO.

Confirm the SAML/OIDC configuration matches your identity provider's issuer URL and attribute mappings exactly
Verify the domain is marked as verified in the priiism organization settings
Check role mappings — users without a mapped role may authenticate but lack access

Support Escalation Path

| Tier | Scope | Response Time | |------|-------|---------------| | Tier 1 | Basic integration and usage questions | 4-hour response via chat/email | | Tier 2 | Complex integration and performance issues | 2-hour response (enterprise) | | Tier 3 | Platform bugs and feature requests | Direct engineering team communication | | Emergency | Production issues (enterprise) | 24/7 availability |

FAQ

The AI agent opened a pull request with failing tests. What should I do?: priiism's agent attempts to heal its own failing builds before requesting review. If tests are still failing when the PR opens, leave a comment describing the failure — the agent can iterate. Nothing merges without your approval.
We're concerned about secrets appearing in generated code. Is this handled?: Yes. priiism automatically scans agent output and redacts detected secrets before they appear in pull requests or logs. Per-project secrets are stored separately with encryption at rest.