tfc-api

CRITICAL DIRECTIVE: Always prefer using the terrapyne CLI (e.g. uv run terrapyne or tfc) for TFC operations. Only fall back to raw API calls (curl/httpie) or the bash scripts listed below if terrapyne is not available in the environment or lacks the required functionality.

Example: Instead of ./scripts/list-runs.sh, use uv run terrapyne run list. Instead of manual curl for workspace discovery, use uv run terrapyne workspace list or tfc project find.

Query TFC workspaces, runs, logs, and team permissions without leaving the terminal.

Authentication

All scripts use a two-tier token resolution: TFC_TOKEN env var takes precedence, falling back to ~/.terraform.d/credentials.tfrc.json. This matters because personal tokens often lack org-level permissions (e.g. team management) that automation tokens have.

# Option 1: Environment variable (preferred for automation/elevated ops)
export TFC_TOKEN=<your-token>

# Option 2: Credentials file (automatic fallback)
# ~/.terraform.d/credentials.tfrc.json — used if TFC_TOKEN is unset

# Typical workflow: source an env file with the automation token
source /tmp/env.sh  # sets TFC_TOKEN

Gotcha: A personal token may return 404 or empty results for team/project operations that an automation token can see. If find-team.sh returns no results but you know teams exist, try an elevated token.

Team & Project Access (permissions management)

Read the reference first: references/tfc-team-project-access.md

Key gotchas:

• Teams API supports q= search — always use it (large orgs may have 1000+ teams)
• Projects API does NOT support q= — must paginate manually via find-project.sh
• tfe_team_project_access with access=admin in Terraform ≠ guaranteed admin in TFC API — always verify
• workspace_access.runs = "read" = users are blocked; must be "apply" for full access

# Full diagnosis + fix workflow
./scripts/find-team.sh {ORGANIZATION} <name_pattern>              # find team IDs
./scripts/find-project.sh {ORGANIZATION} <name_pattern>           # find project ID
./scripts/get-team-project-access.sh <project_id> <team_id>       # current perms
./scripts/compare-team-project-access.sh <proj_a> <team_a> \
                                         <proj_b> <team_b>        # diff vs reference
./scripts/set-team-project-access.sh <project_id> <team_id> admin # fix

Workspace Discovery

Prefer terrapyne:

tfc workspace list                            # list all workspaces
tfc workspace show                            # show current workspace (auto-detected)
tfc project find <pattern>                    # find projects by name fragment

Script/curl fallback (when terrapyne unavailable or lacks feature):

# Search by name pattern (fuzzy)
BASE="https://app.terraform.io/api/v2"
curl -s -H "Authorization: Bearer $TFC_TOKEN" \
  "$BASE/organizations/{ORGANIZATION}/workspaces?search[name]=<pattern>" \
  | jq '.data[] | {id, name: .attributes.name}'

# Search by wildcard name pattern (prefix, suffix, or substring)
# Note: square brackets should be percent-encoded in URLs: [ as %5B and ] as %5D
curl -s -H "Authorization: Bearer $TFC_TOKEN" \
  "$BASE/organizations/{ORGANIZATION}/workspaces?search%5Bwildcard-name%5D=*<pattern>*" \
  | jq '.data[] | {id, name: .attributes.name}'

# Using httpie (modern replacement for curl)
# httpie handles encoding and provides cleaner syntax
http GET "$BASE/organizations/{ORGANIZATION}/workspaces" \
  "Authorization: Bearer $TFC_TOKEN" \
  "search[wildcard-name]==*<pattern>*" \
  | jq '.data[] | {id, name: .attributes.name}'

# Get workspace by exact name

curl -s -H "Authorization: Bearer $TFC_TOKEN"
"$BASE/organizations/{ORGANIZATION}/workspaces/<workspace-name>"
| jq '{id: .data.id, name: .data.attributes.name}'

## Run Management

**Prefer terrapyne:**
```bash
tfc run list -w <workspace-name>
tfc run trigger <workspace-name> -m "message" --wait
tfc run watch <run-id>
tfc run apply <run-id> -m "Approving"
tfc run logs <run-id> --stage plan
tfc run logs <run-id> --stage apply
tfc run discard <run-id> -m "reason"

Script fallback (when terrapyne unavailable or lacks feature):

# List recent runs
./scripts/list-runs.sh {ORGANIZATION} <workspace-name>

# Trigger a run
./scripts/trigger-run.sh {ORGANIZATION} <workspace-name> false "message"

# Wait for completion
./scripts/wait-for-run.sh <run-id>

# Approve a planned run
./scripts/apply-run.sh <run-id> "Approving"

Unlocking a workspace blocked by old runs

Prefer terrapyne where possible:

tfc run discard <run-id> -m "Discarding to re-plan"
# Note: run cancel (for pending/planning runs) is not yet in terrapyne (tracked as F14)
# Use curl fallback below for cancel

Curl fallback:

TFC_TOKEN=$(jq -r '.credentials."app.terraform.io".token' ~/.terraform.d/credentials.tfrc.json)
BASE="https://app.terraform.io/api/v2"

# Discard a planned/cost_estimated run
curl -s -X POST -H "Authorization: Bearer $TFC_TOKEN" \
  -H "Content-Type: application/vnd.api+json" \
  -d '{"comment":"Discarding to re-plan"}' \
  "$BASE/runs/<run-id>/actions/discard"

# Cancel a pending/planning run (discard won't work on pending)
curl -s -X POST -H "Authorization: Bearer $TFC_TOKEN" \
  -H "Content-Type: application/vnd.api+json" \
  -d '{"comment":"Cancelling"}' \
  "$BASE/runs/<run-id>/actions/cancel"

Gotcha: If the workspace is locked by run A and run B is queued behind it, you must discard/cancel A first — B won't start until A releases the lock. Check locked-by on the workspace to find the blocking run:

curl -s -H "Authorization: Bearer $TFC_TOKEN" \
  "$BASE/organizations/{ORG}/workspaces/<name>" \
  | jq '.data.relationships["locked-by"].data.id'

Switching VCS branch for speculative plans

Note: terrapyne does not yet have a workspace set-branch command (tracked as F13). Use the script fallback:

./scripts/set-workspace-branch.sh {ORGANIZATION} <workspace-name> <branch>

Useful for testing a consolidated branch (iac-reveng) against workspaces that normally point at per-env branches (iac-reveng-dev). Remember to reset the branch after validation.

Workspace State Inspection

Prefer terrapyne:

tfc run logs <run-id> --stage plan
tfc run logs <run-id> --stage apply

Script fallback:

# Get plan/apply logs
./scripts/get-plan.sh <run-id>
./scripts/get-apply.sh <run-id>

Helper Scripts

| Script | Purpose | |--------|---------| | find-workspace.sh | Find workspaces by name or wildcard pattern (prefix/suffix/substring) | | find-team.sh | Find teams by name pattern using q= (handles 1000+ teams) | | find-project.sh | Find projects by name fragment (paginates all pages) | | list-team-project-access.sh | List all team access entries for a project | | get-team-project-access.sh | Get one team's access record for a project | | compare-team-project-access.sh | Diff permissions between two teams | | set-team-project-access.sh | Update team access level on a project | | list-runs.sh | List recent runs for a workspace | | trigger-run.sh | Trigger a new run (supports destroy) | | apply-run.sh | Confirm/apply a planned run | | wait-for-run.sh | Poll status until terminal state | | get-plan.sh | Download and parse plan logs | | get-apply.sh | Download and parse apply logs | | get-workspace-info.sh | Detect ORG and Workspace from local TF config | | set-workspace-branch.sh | Update VCS branch for speculative plans | | open-workspace.sh | Open TFC workspace in browser | | publish-registry-module-vcs.sh | Publish a registry module via VCS | | check-registry-module-status.sh | Check registry module status | | delete-registry-module.sh | Delete a registry module |

References

• references/tfc-team-project-access.md — Team permissions CRUD, pagination gotchas, access level semantics
• references/tfc-api-reference.md — Workspace/run/plan/log operations
• references/tfc-vcs-validation.md — VCS configuration validation
• TFC API Docs