ai-skills
AGENTS.md36 starsChanged 3 months agoMITView on GitHub
17 skills covering the Infisical platform. Read the router below before loading a skill — several skills cover adjacent concepts that are easy to confuse, and loading the wrong one produces confidently wrong answers. Check these before answering. Each pair looks similar and routes differently. Several features depend on another being set up first. Check the dependency before debugging.
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,542 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository17 other files it gives its agents.
- infisical-access-controlSkill
- infisical-agentSkill
- infisical-apiSkill
- infisical-app-connectionsSkill
- infisical-dynamic-secretsSkill
- infisical-gatewaySkill
- infisical-kmsSkill
- infisical-kubernetes-operatorSkill
- infisical-pamSkill
- infisical-pkiSkill
- infisical-secret-rotationSkill
- infisical-secret-scanningSkill
- infisical-secret-syncsSkill
- infisical-self-hostSkill
- infisical-setupSkill
- infisical-ssoSkill
- infisical-terraformSkill
The file
We split this AGENTS.md into the rules it gives an agent
5 rules. About 1 is common, 3 are distinctive, and 1 is very specific to this project.
≈1
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 31
Check these before answering.
- Common: line 89
All 83 connection types and their exact auth methods, platform-managed credentials, Gateway-routed connections, connection-level credential …
- Distinctive: line 139
When re-verifying, read the code, not the docs prose.
- Very specific: line 164
Do not edit references/api-surface.md or references/credentials-by-connection.md by hand — they carry a generated-file banner and the next …
- Distinctive: line 167
Prefer extending this pattern over hand-maintaining any new enumerable fact.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Infisical AI Skills
17 skills covering the Infisical platform. Read the router below before loading a skill — several skills cover adjacent concepts that are easy to confuse, and loading the wrong one produces confidently wrong answers.
Router: pick a skill from the user's intent
| The user wants to... | Skill |
| Get secrets into an app, container, CLI, CI/CD, or via an SDK | infisical-setup |
| Call the REST API directly | infisical-api |
| Use Terraform / HCL | infisical-terraform |
| Render secrets to a file, or run a sidecar/init container | infisical-agent |
| Sync secrets into Kubernetes Secrets/ConfigMaps | infisical-kubernetes-operator |
| Push secrets out to a third-party service | infisical-secret-syncs |
| Authenticate Infisical out to a third party | infisical-app-connections |
| Get a brand-new short-lived credential per request | infisical-dynamic-secrets |
| Change an existing credential on a schedule | infisical-secret-rotation |
| Let a human or AI agent reach infrastructure, recorded, without a credential | infisical-pam |
| Issue or manage X.509 / TLS certificates, or sign code | infisical-pki |
| Encrypt / sign data with a managed key | infisical-kms |
| Find leaked credentials in source code | infisical-secret-scanning |
| Configure human login (SAML/OIDC/LDAP) or SCIM provisioning | infisical-sso |
| Define roles, permissions, approvals, or stream audit logs | infisical-access-control |
| Reach a resource with no public endpoint | infisical-gateway |
| Deploy Infisical itself | infisical-self-host |
Disambiguation: the pairs that get confused
Check these before answering. Each pair looks similar and routes differently.
| Sounds like | Actually two things |
| "temporary credentials" | Dynamic secrets create a new credential per lease. Secret rotation changes one existing credential on a timer. New vs existing. |
| "give someone database access" | PAM brokers a recorded session and never reveals the credential. Dynamic secrets hand a real credential to a workload. Human/agent vs application. |
| "connect Infisical to AWS" | App Connection = Infisical authenticating outward. Machine identity = a workload authenticating inward. Check the direction. |
| "OIDC" / "LDAP" | infisical-sso for humans logging in. infisical-setup for machine identity auth. Also an LDAP App Connection and LDAP dynamic secrets. Four unrelated things. |
| "certificates" | infisical-pki for X.509/TLS. infisical-dynamic-secrets for SSH certificates. SSH is not in the PKI product. |
| "sync" | Secret Syncs push secrets. PKI Syncs push certificates — a separate feature in infisical-pki. |
| "install the Helm chart" | secrets-operator chart = the operator (infisical-kubernetes-operator). The standalone chart = the platform (infisical-self-host). |
| "rotation" | Secret Rotation writes new values into secrets. App Connection credential rotation refreshes Infisical's own stored credential and writes nothing. PAM account rotation is a third thing. |
| "approval" | Secrets → infisical-access-control. Infrastructure sessions → infisical-pam. Certificate issuance or code signing → infisical-pki. |
| "encryption key" | infisical-kms to hold a key and do the crypto. infisical-setup if they just want to store key material as a secret. |
Shared prerequisites
Several features depend on another being set up first. Check the dependency before debugging.
| Feature | Requires |
| Secret Sync | An App Connection of the matching type |
| Secret Rotation | An App Connection of the matching type |
| PKI external CA / PKI Sync | An App Connection, usually plus a Gateway |
| Secret Scanning data source | An App Connection — github-radar, not github, for GitHub |
| Anything reaching a private network | A Gateway (only 16 connection types accept one) |
| SCIM provisioning | Email domain verification and SAML/OIDC already configured |
| Any SSO login | Email domain verification |
| HSM-backed keys | A Gateway in the HSM's network |
Skills
infisical-setup
Getting secrets into applications and platforms. CLI, all 9 SDKs (Node.js, Python, Go, Java, .NET, Ruby, PHP, Rust, C++), Docker, CI/CD, and all 13 machine identity auth methods.
infisical-api
The REST API: /api/v4/secrets CRUD and batch operations, projects, identities, all 13 auth-method
login endpoints, which endpoints paginate and which do not, rate limits.
infisical-terraform
The Infisical Terraform provider: the nested auth = { universal | oidc } block, ephemeral resources
keyed by name, data sources, project roles, Terraform Cloud OIDC.
infisical-agent
The Agent daemon: YAML config, 6 agent auth methods, Go template functions including the
SSH-required principals argument, sinks, polling, on-change commands, caching.
infisical-kubernetes-operator
The operator: v1beta1 CRDs (InfisicalConnection, InfisicalAuth, InfisicalStaticSecret) and
legacy v1alpha1 (InfisicalSecret, InfisicalDynamicSecret, InfisicalPushSecret), Helm install,
auto-reload, templating.
infisical-secret-syncs
Pushing secrets to all 48 destinations. App Connections, key schemas, initial-sync enum values, mapping behavior (AWS Secrets Manager only), provider quirks.
infisical-app-connections
All 83 connection types and their exact auth methods, platform-managed credentials, Gateway-routed connections, connection-level credential rotation. The shared prerequisite for syncs, rotations, PKI, and scanning.
infisical-dynamic-secrets
On-demand short-lived credentials across all 30 providers. Lease lifecycle, TTLs, creation statements, SSH certificates, Gateway setup for private networks.
infisical-secret-rotation
Rotating existing credentials across 28 providers. Dual-phase vs single-phase, the two-user pattern
for SQL, rotationInterval and rotateAtUtc, parameters and secretsMapping.
infisical-pam
Brokered privileged access for humans and AI agents across 13 account types. Accounts, folders, templates, memberships, session recording, just-in-time approvals, agentic access.
infisical-pki
Certificate Management: 9 CA types, Policies/Profiles/Applications, 4 enrollment methods (API/ACME/EST/SCEP), 12 PKI Syncs, code signing, HSM connectors, post-quantum algorithms.
infisical-kms
Managed cryptographic keys: the 3 key usages, symmetric/asymmetric/HMAC algorithms, encrypt, decrypt, sign, verify, MAC, rotation, external KMS, KMIP, cosign.
infisical-secret-scanning
Detecting leaked credentials: 3 cloud data sources, full vs diff scans, finding lifecycle, the
infisical scan CLI and pre-commit hooks, noise reduction, AWS honey tokens.
infisical-access-control
Governance: org and project roles, custom roles, the granular secret actions that separate
describeSecret from readValue, ABAC, temporary access, approval policies, audit log streams.
infisical-sso
Human login and provisioning: SAML/OIDC/LDAP plus free Google/GitHub SSO, SCIM, group-to-role
mapping, SSO enforcement and the /login/admin break-glass portal.
infisical-gateway
Reaching private networks: the gateway/relay architecture, exact ports and egress rules, deployment, Gateway Pools for HA, and which features can be routed through one.
infisical-self-host
Deploying Infisical: Docker, Docker Compose, Kubernetes Helm, environment variables, Redis
requirements including the mandatory noeviction policy, FIPS 140-3, scaling and HA.
Maintaining these skills
Skill content is a vendored snapshot of a moving codebase. An audit found that a drifted skill
performs worse than no skill at all — stale specifics override correct model knowledge. See
evals/accuracy-audit-2026-08/.
When re-verifying, read the code, not the docs prose. Counts and enum values live in:
| Fact | Source |
| Secret syncs | backend/src/services/secret-sync/secret-sync-enums.ts |
| Secret rotations | backend/src/ee/services/secret-rotation-v2/secret-rotation-v2-enums.ts |
| Dynamic secret providers | backend/src/ee/services/dynamic-secret/providers/models.ts |
| App Connections | backend/src/services/app-connection/app-connection-enums.ts |
| Machine identity auth methods | backend/src/db/schemas/models.ts (IdentityAuthMethod) |
| PAM account types | backend/src/ee/services/pam/pam-enums.ts |
| CA types | backend/src/services/certificate-authority/certificate-authority-enums.ts |
| PKI syncs | backend/src/services/pki-sync/pki-sync-enums.ts |
| KMS algorithms | backend/src/lib/crypto/{cipher,sign,hmac}/types.ts |
| Permissions | backend/src/ee/services/permission/{org,project}-permission.ts |
| API routes | backend/src/server/routes/ |
Generated files
The App Connection factual references are generated, not hand-written:
python3 tools/generate-app-connection-refs.py # regenerate from the Infisical repo
python3 tools/generate-app-connection-refs.py --check # fail if the checked-in output is staleDo not edit references/api-surface.md or references/credentials-by-connection.md by hand — they
carry a generated-file banner and the next run overwrites them. Change the generator instead.
Prefer extending this pattern over hand-maintaining any new enumerable fact.
Wrapper drift
plugins/<name>/skills/<name>/ is a copy of skills/<name>/. They drift silently. CI checks
this, and so should you before committing.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.