Secret Manager Autonomy Journal

VerifiedSafe

Appends a structured entry to an audit journal for every autonomous GCP Secret Manager action, without logging secret values.

Sby Skills Guide Bot
SecurityBeginner
7/27/2026
Claude Code
#audit-trail#secrets-management#gcp#git-commit#documentation

Recommended for

Automated summary

Appends a structured entry to docs/secret-manager-journal.md for every autonomous GCP Secret Manager action, successful or failed, without ever recording secret values.

Strengths

  • Enforces a fixed field table (repo, GCP project, agent, session, trigger, action type, mechanism, result, ADR) so audit entries stay consistent across projects
  • Append-only journal with per-project sections and newest-first ordering
  • Explicit safety rules: no secret values, no edits to past entries, and failures are logged just like successes
  • Automatically commits the entry with a normalized git message

Limitations

  • Relies on CLAUDE.md and the live conversation to fill fields; it must ask the user when something cannot be determined
  • Does not replace GCP Cloud Audit Logs — it captures agent context (who decided, why, which session), not raw API calls
  • Experimental maturity with no automated validation of entry format
When to use it

Use it right after an autonomous Secret Manager operation (shell creation, version add, IAM binding, import, delete) to keep an auditable trail.

When not to use it

Skip it for plain API calls already covered by Cloud Audit Logs, or when no usable session context exists to complete the required fields.

Security analysis

Safe
Quality score86/100

The skill only appends audit metadata to a single documentation file and explicitly forbids writing secret values, modifying prior entries, or touching the skill file itself. No network calls, no shell beyond scoped git add/commit, and no obfuscated or destructive operations. The breadth of Write/Edit and the auto-commit are minor, mitigated by clear safety rules.

Findings
  • •Write/Edit allowed-tools are not path-scoped; the skill relies on instructions to only touch docs/secret-manager-journal.md
  • •Performs git add/commit automatically without human confirmation (limited to the journal file)
How this analysis is made

Examples

Log a successful secret creation
I just created the secret `db-password` in GCP project acme-prod via Terraform (workflow deploy-infra run #482). Add an entry to the Secret Manager journal.
Log a failed IAM binding
Adding the IAM binding for service account `ci-runner@acme-prod.iam.gserviceaccount.com` on secret `api-token` failed with permission denied using the gcloud CLI. Record this failure in the journal.
Journal a manual import
/secret-manager-journal — document the manual import of the `stripe-key` secret I did in this session.

name: secret-manager-journal description: "Appends an entry to docs/secret-manager-journal.md recording every autonomous GCP Secret Manager action: repo, project, date, session ID, mechanism, and result. Use when an agent completes or fails a Secret Manager operation and needs to document it." allowed-tools:

  • Read
  • Edit
  • Write
  • Bash(git add *)
  • Bash(git commit *) maturity: experimental source-experiment: core evidence: "First use 2026-04-20 — project-life-130."

Secret Manager Autonomy Journal

Role

You are an Audit Recorder. You append a structured entry to docs/secret-manager-journal.md documenting every autonomous GCP Secret Manager action — successes and failures — across all projects. You never write to the skill file itself and you never log secret values.

Context — Read First

Before writing, read in parallel:

  • docs/secret-manager-journal.md — check if file and project section exist
  • CLAUDE.md — extract repo name and GCP_PROJECT_ID

Instructions

Step 1: Extract action details from context

From the current conversation, extract:

| Field | Source | |-------|--------| | repo | git remote or CLAUDE.md | | gcp_project | CLAUDE.md GCP Bootstrap section | | date | today's date (YYYY-MM-DD) | | agent | current model ID (e.g., claude-sonnet-4-6) | | session_id | workflow run ID if available, else session-<YYYYMMDD-HHMMSS> | | trigger | workflow name + run ID, or "manual session" | | action_type | one of: create-shell / add-version / iam-binding / import / delete | | secrets_affected | comma-separated secret names — NO values | | mechanism | Terraform / gcloud CLI / GitHub Actions workflow / API call | | result | ✅ Success or ❌ Failed — reason | | adr_reference | ADR number if applicable, else "none" | | details | 1–3 bullet points: what was done and why — NO secret values |

If any field cannot be determined from context, ask the user before writing.

Step 2: Ensure file and project section exist

If docs/secret-manager-journal.md does not exist, create it with this header:

# Secret Manager Autonomy Journal

Cross-project audit trail for all autonomous GCP Secret Manager actions.
Append-only. Never delete or modify previous entries.

---

If the project section ## Project Journal — <repo> does not exist, append it:

## Project Journal — `<owner>/<repo>`

**GCP Project:** `<gcp_project>`
**Domain:** `or-infra.com`

---

Step 3: Append the new entry

Add the entry at the top of the correct project section (newest first), immediately after the --- separator that follows the section header.

Use this exact format:

### [<date>] <action_type> — <short description>

| Field | Value |
|-------|-------|
| **Repo** | `<owner>/<repo>` |
| **GCP Project** | `<gcp_project>` |
| **Date** | <date> |
| **Agent** | <agent> |
| **Session ID** | <session_id> |
| **Trigger** | <trigger> |
| **Action type** | <action_type> |
| **Secrets affected** | <secret names — NO values> |
| **Mechanism** | <mechanism> |
| **Result** | <✅ Success / ❌ Failed — reason> |
| **ADR reference** | <adr_reference> |

**Details:**
- <bullet 1>
- <bullet 2>

---

Step 4: Commit

git add docs/secret-manager-journal.md
git commit -m "docs: log secret manager action — <repo> (<action_type>)"

Safety Rules

  1. NEVER write to the skill file itself — only to docs/secret-manager-journal.md.
  2. NEVER include secret values in any field, including Details bullets — write "[omitted — security policy]" if a value appears in context.
  3. NEVER skip logging a failed action — ❌ entries are as important as ✅ entries.
  4. NEVER modify previous entries — append only; treat the file as immutable history.
  5. NEVER duplicate a GCP Cloud Audit Log entry — this journal records agent context (who decided, why, which session), not raw API calls.

Examples

User: /secret-manager-journal (after bootstrap.yml stored RAILWAY_PROJECT_TOKEN)

Agent behaviour: Reads CLAUDE.md → repo=edri2or/project-life-130, gcp=project-life-130. Reads journal file → project section exists. Extracts: action_type=add-version, secrets=RAILWAY_PROJECT_TOKEN, mechanism=gcloud secrets versions add --data-file=-, trigger=bootstrap.yml run ID: 24670020337, result=✅ Success. Appends entry at top of project section. Commits.

User: /secret-manager-journal (Terraform apply failed — Error 409)

Agent behaviour: Result field: ❌ Failed — Error 409: Secret CLOUDFLARE_ADDITIONAL_UZER_TOKEN already exists. Action type: create-shell. Details: "Terraform attempted to create secret that already existed out-of-band. Fix: add import block to secrets.tf (ADR 0008)." Logs the failure entry — does NOT skip because it was a failure.

User: /secret-manager-journal (new project, first action)

Agent behaviour: Reads journal file — finds no section for edri2or/my-new-project. Creates new ## Project Journal — edri2or/my-new-project section with GCP project header. Appends first entry under it. Commits.

From the file to a first result

The download contains exactly the content displayed on this page. This copy may differ from the latest version in the source repository.

View the source and supporting files

This download contains SKILL.md only. Any scripts, references, templates and dependencies are not included. Get the required files from the source, preserving their directory structure, before continuing.

1. Download the file

Download SKILL.md

From the root of a test project, save the file as .skill-downloads/secret-manager-autonomy-journal-ms2s7cse/SKILL.md, or use the command below. This folder is for inspecting the skill before installation.

Command for macOS/Linux or WSL: requires a POSIX shell and curl. On Windows without WSL, use the download button and create the folder manually.

mkdir -p '.skill-downloads/secret-manager-autonomy-journal-ms2s7cse' && curl --fail --show-error --location 'https://skills-guides.com/api/skills/secret-manager-autonomy-journal-ms2s7cse/download' --output '.skill-downloads/secret-manager-autonomy-journal-ms2s7cse/SKILL.md.part' && mv '.skill-downloads/secret-manager-autonomy-journal-ms2s7cse/SKILL.md.part' '.skill-downloads/secret-manager-autonomy-journal-ms2s7cse/SKILL.md'

2. Check prerequisites

Read SKILL.md before using it. If it references scripts or external services, inspect them and prepare the required access in your test environment.

No recognized prerequisites section in this file. This does not mean it is self-contained: libraries, tools and access requirements may be mentioned elsewhere in the content or source repository.

The platforms on this page are catalog indications. The paths below follow tool documentation; they do not prove this skill has been run in those tools. No minimum tool version is confirmed for this listing.

3. Load the skill

After inspection, copy the complete folder into your project using one of the commands below, then open a new tool session. The command stops if a folder with the same name already exists: compare the files before replacing it.

Claude Code
test -f '.skill-downloads/secret-manager-autonomy-journal-ms2s7cse/SKILL.md' && test -s '.skill-downloads/secret-manager-autonomy-journal-ms2s7cse/SKILL.md' && mkdir -p '.claude/skills' && test ! -e '.claude/skills/secret-manager-journal' && cp -R '.skill-downloads/secret-manager-autonomy-journal-ms2s7cse' '.claude/skills/secret-manager-journal'

OpenSkills

OpenSkills must already be installed. The local install/read procedure is checked with OpenSkills 1.5.0, which requires Node.js ≥ 20.6.0. Catalog slugs are not resolved automatically.

Official OpenSkills installation
openskills install './.skill-downloads/secret-manager-autonomy-journal-ms2s7cse'

openskills read 'secret-manager-autonomy-journal-ms2s7cse'

Reading should display the installed file content and its base directory. It does not run the skill or install its dependencies.

For an agent that reads AGENTS.md, then sync the index. This command changes AGENTS.md: review the diff before using it.

openskills sync

Manual method or another tool: attach SKILL.md to a conversation, or open it in an agent that can read files. Include any required supporting resources. No universal folder is recognized by every tool.

4. Try a first task

Use a short, synthetic input: an API endpoint, a small document or an interface brief. In a new conversation, attach the file or give the agent its path, then adapt this request:

Read .skill-downloads/secret-manager-autonomy-journal-ms2s7cse/SKILL.md and apply its instructions to this example: [describe a test input]. Identify any missing files, dependencies or access. Produce the result requested by the skill without inventing data or running unreviewed scripts.

Expected result: output matching the skill instructions, or a precise list of missing prerequisites. A read confirmation alone does not validate a first use. If an external resource is unavailable, that path remains untested.

Related skills