Use when the agent needs to add a comment to an `np` (nitpicking) issue — to record reasoning, capture an investigation finding, leave a handoff note, document a trade-off considered, or annotate any issue (including closed ones). Triggers on prompts like "add a comment to FOO-a3bxr saying ...", "leave a note on the issue", "record that we tried X", "comment on FOO-12345 that ...". Comments do not require a claim.
Use when the agent needs to file a new `np` (nitpicking) issue — an incidental discovery while working on something else, a child task or sub-epic when decomposing a parent epic, a deferred follow-up, or any new task or epic the user asks to be created. Triggers on prompts like "create an issue for ...", "file a follow-up", "decompose this epic into ...", "add a child task under FOO-c4npt".
Use when the agent needs to pick up the next ready issue from the `np` (nitpicking) tracker, claim a specific issue by ID, browse the ready queue, or filter ready work by label or role (e.g., "claim the next ready bug", "pick up the next task in area:auth", "claim FOO-a3bxr"). Does not cover reading the claimed issue's content or finishing work.
Use when the agent has a claimed `np` (nitpicking) issue and needs to close it as complete, release the claim without changing the issue's state, or defer the issue for later. Triggers on prompts like "close the issue", "mark it done", "release the claim", "defer this for now". Does not cover reopening or undeferring (escape-hatch via `np-help-discipline`); commenting is handled separately by `np-commenting`.
Use when about to invoke an `np` (nitpicking) command, subcommand, or flag that has not been verified in the current session, or when unsure whether a flag or subcommand exists. Also use as the escape hatch for any `np` activity not covered by another skill (e.g., `np issue reopen`, `np issue undefer`, `np epic close-completed`, `np admin backup`, `np agent name`, `np json update`).
Use when the agent needs to add or remove labels on an existing `np` (nitpicking) issue (often across multiple issues at once), or inspect the workspace's label vocabulary for drift or to ground a new label key. Triggers on prompts like "add foo:bar to all issues about X", "remove the kind:bug label from FOO-a3bxr", "what labels are in use", "tag these issues with sprint:2026-w17". Does not cover label-based filtering of the ready queue (use `np-finding-work`) or label-based filtering during issue lookup (use `np-reading-issues`).
Use when the agent needs to express, remove, or list a dependency or informational link between two `np` (nitpicking) issues — adding `blocked_by` because one issue cannot proceed until another finishes, adding `refs` for a "see also" pointer, removing a relationship that no longer applies, or listing the workspace's relationships at a glance. Triggers on prompts like "mark FOO-a3bxr as blocked by FOO-c4npt", "FOO-12345 references FOO-67890", "remove the blocker", "show me the blocking chains".
Use when the agent needs to understand a specific `np` (nitpicking) issue (typically right after claiming it), explore the issue's parent, siblings, or `refs` neighbours for context, look up an issue by ID, view its history, search for issues by text, or list issues by label or other filters. Triggers on prompts like "what does FOO-a3bxr cover", "what do I need to do for this issue", "do we already have an issue for X", "show me the history of FOO-a3bxr".