Imported from Sidewatch/swift-agent-session (AGENTS.md). Install upstream with npx skills add Sidewatch/swift-agent-session. Copyright stays with the author.

Swift Agent Session

A tiny, dependency-free reader for terminal AI coding-agent session transcripts. It maps an agent's on-disk session onto one agent-agnostic model — an activity timeline, token/cost telemetry, and an edited-files/to-dos roll-up — so review surfaces stay identical across agents. Read-only: it never talks to a model, keeps no account, and sends no telemetry.

• Module AgentSession in Sources/AgentSession; tests in Tests; swift test is the whole check.
• Swift 6 language mode, tools 6.2, macOS 14+, no dependencies unless the README says so.
• Part of the Sidewatch package family; every package follows the same layout and PR rules.

Module map

• Adapters/ — the engine: adapters: Agents, ClaudeCodeAdapter
• Models/ — value types — the shape of a thing, nothing else: AgentSummary, AgentUsage, ClaudeCredentials, ClaudeQuota, TimelineEvent, TurnBoundary, UsageReport
• Protocols/ — protocols the module exposes: AgentAdapter
• Transcripts/ — the engine: transcripts: SessionMemo, TranscriptCache, TranscriptState
• Usage/ — the engine: usage: ClaudeQuotaCache, ClaudeUsageEndpoint, ModelPricing, UsageAggregator, UsageRecord, UsageTotals

Rules

Read CONTRIBUTING.md before changing anything: it is the layout and PR rulebook for this package.

• Auditing? Read AUDIT.md first — what the last full audit checked and fixed, and the known non-issues to skip; extend it, do not redo it.