Automatically recover working context after session compaction or when continuation is implied but context is missing. Works across Discord, Slack, Telegram,...
Security Analysis
high confidenceThe skill's requested actions (read channel history, synthesize a context summary, ask before persisting) align with its stated purpose and it does not ask for extra credentials or install code, but it will read message history and can auto-trigger so review connector permissions and privacy settings before enabling.
Name/description (recover missing conversation context) matches the instructions: fetch channel/thread history, parse messages, and synthesize a summary. There are no unrelated environment variables, binaries, or install steps requested.
Instructions confine operations to platform/session APIs (message:read, sessions_list, sessions_history) and explicitly prohibit broad filesystem scans and external exfiltration. They do parse for identifiers like URLs, branch names, PRs and may look for 'working directories' mentioned in messages (parsing only). Note: the skill can auto-trigger on compaction indicators and will read channel/thread history automatically when triggered, which has privacy implications depending on connector permissions.
Instruction-only skill with no install spec and no code files. This minimizes disk/exec risk — nothing is downloaded or written by default.
The skill declares no required env vars or credentials (proportional). However, its operation depends on the agent's channel connectors having message-read permissions (e.g., Slack/Discord tokens). Those connector permissions can grant broad access to chat history, so review platform scopes/credentials before use.
always:false and user-invocable:true. The skill may auto-trigger in-session (appropriate for its purpose) but does not request permanent presence or modify other skills. It requires explicit consent before writing recovered context to disk or persistent memory.
Guidance
This skill appears coherent and does what it says: it reads channel/thread history and synthesizes a recovered context, then asks before saving anything. Before installing, verify the following: 1) The agent connectors (Slack, Discord, Telegram, etc.) only have the minimal scopes you intend — message-read scopes can expose a lot of history. 2) Confirm your runtime actually enforces the 'no filesystem scans / no external sending' constraints described in SKILL.md (instruction-only skills rely on the agent runtime to honor guidance). 3) If you prefer tighter control, disable automatic invocation or require manual invocation, and ensure the skill prompts you before persisting recovered context. If you want extra assurance, run the skill in a limited-test workspace first to observe behavior.
Latest Release
v1.2.2
Scanner hardening: add provenance fields, align _meta version, keep API-only recovery scope and consent-based persistence
More by @jdrhyne
Published by @jdrhyne on ClawHub
