Worksync

worksync — log a sync-worthy entry

Appends a single timestamped bullet to today's note at ${WORKSYNC_VAULT_DIR:-$HOME/worksync}/YYYY-MM-DD.md via the bundled helper script ${CLAUDE_PLUGIN_ROOT}/scripts/worksync-append.sh.

The file's purpose: a low-friction daily log of things the user will want to mention at a work sync, standup, or weekly review. Phrase entries for a technical peer audience.

Arguments

The skill accepts a free-form argument from $ARGUMENTS:

• If non-empty: treat it as the message, with optional leading type hint.
• If empty: derive message from recent session activity (see "Inferring content" below).

Supported types (passed as first arg to the helper):

| Type | When to use | |--------------|--------------------------------------------------------------------------| | research | A question was investigated and produced a concrete finding or decision. | | ops | A deploy, rollout, migration, or infrastructure operation finished. | | checkpoint | User-requested mid-task marker ("log this", "checkpoint"). | | note | Anything else worth mentioning that doesn't fit the above. |

Types pr, issue, merge are reserved for the automatic hook — do not emit them from the skill. If the user explicitly asks to log one manually, use note and include the URL in the message.

Inferring content (no arguments case)

When invoked with no arguments, scan the most recent turns of the current session and produce ONE entry summarizing the most recent sync-worthy event. Do not produce multiple entries in one invocation — call the skill again if more is needed.

Selection priority (pick the most recent qualifying item):

1. An ops action that visibly succeeded (deploy, apply, rollout, migration,

infrastructure reconfiguration, provisioning run against real hosts).

1. A research conclusion — user asked a question, you investigated, you

reached a definite answer or recommendation.

1. A checkpoint the user explicitly requested.

Skip if the most recent activity was just reading/exploring without producing a decision or action — nothing to log yet.

Message format

Keep the message to one line, English, technical-peer tone.

Good examples:

• research — confirmed HTTP/2 keepalive is the right fix for the connection-reset issue; ADR pending
• ops — rolled out v1.42.3 to staging, all replicas healthy inside 90s
• ops — ran db migration 0042 on production, 0 rows skipped, 12s runtime
• research — decided to drop Redis cluster mode — single-primary + replica covers our load

Bad examples (do not emit):

• checked pod status (not sync-worthy)
• discussed architecture (no outcome)
• anything with customer names, credentials, or internal-only identifiers

Execution

Run the helper script. Do not write to the file directly — the helper handles directory creation, date-based filenames, and deduplication.

"${CLAUDE_PLUGIN_ROOT}/scripts/worksync-append.sh"  ""

Exit on failure is fine — surface the error to the user. On success, reply with one short line: which file was updated and what was logged.

Redaction rules

Before appending, strip or rewrite:

• Customer, client, or project names → generic labels ("production cluster",

"test environment", "a customer request").

• Internal hostnames or cluster identifiers that reveal topology → generic.
• Credentials, tokens, private IP addresses, internal URLs.

If the event cannot be described without private details, log a vague version ("reviewed an internal auth change") rather than skipping — the user still wants a timestamp marker they can expand on verbally at the sync.

Configuration

Users can point the note directory at any location by exporting WORKSYNC_VAULT_DIR in their shell, Claude settings.json env block, or a project .env. Common targets:

• ~/Documents/Obsidian/MyVault/worksync — Obsidian vault subfolder
• ~/Library/Mobile Documents/iCloud~md~obsidian/Documents//worksync — macOS iCloud-synced Obsidian vault
• ~/logseq/journals/worksync — Logseq graph
• ~/worksync — the default, plain folder

Daily files are plain markdown, so any editor will do.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: lexfrei
• Source: lexfrei/ccc
• License: BSD-3-Clause

Install and usage instructions live in the source repository linked above.