v2.1.0

What's Changed

• fix(feature): --feature dropped every definition and reported the project's numbers by @Shahid-BugB in #22
• Prune internal planning docs, point the rest at docs.bugb.io by @Animesh-Sri-bugb in #23
• fix(templates): give the starved annotation verbs a trigger, not just permission by @Shahid-BugB in #24
• fix: stop showing agents a truncated model and calling it complete by @Shahid-BugB in #25
• guardlink paths — ask the flow graph where data reaches without a control by @Shahid-BugB in #26
• feat(verify): stale claim detection — bind every claim to its code and notice when it changes by @Shahid-BugB in #27
• feat(blame,dashboard): who wrote each risk, and a dashboard that directs instead of reporting by @Shahid-BugB in #30
• fix(parser): read the doc-comment forms developers write, and say so when we cannot by @Animesh-Sri-bugb in #31
• Make a GuardLink acceptance cost something, and make guardlink ci a gate a team can adopt by @Animesh-Sri-bugb in #32
• feat(playbooks,gate,hypothesis): the tool owns the method and the evidence bar, gates what an agent wrote, and records what happened when an exposure was tested by @Shahid-BugB in #33
• fix(diagrams): a render budget, so an undrawable diagram is never written or certified as fine by @Animesh-Sri-bugb in #34
• Query-driven graph views: ask a question, get an answer sized to be read by @Animesh-Sri-bugb in #35
• feat(hypothesis): a ledger for what happened when an exposure was tested (GAP-51) by @Shahid-BugB in #36
• fix(acceptance): an acceptance no longer asserts more than anyone checked by @Animesh-Sri-bugb in #37
• fix(hypothesis): join scan findings by claim key, not by whatever sits on the tested line by @Shahid-BugB in #40
• fix(parser): read annotations in every language §2.9 claims, not six of them by @Animesh-Sri-bugb in #38
• fix(merge,ci): a command that read nothing no longer reports a clean estate by @Animesh-Sri-bugb in #39
• fix(merge,ci): scope the cross-repo join by owner, and stop answering the estate question one repo at a time by @Animesh-Sri-bugb in #41
• fix(ci,init): say what the model knows — lapsing acceptances, a coverage floor, and no threat model we do not have by @Animesh-Sri-bugb in #42
• fix(parser,init): say when annotations are written where nothing reads them by @Animesh-Sri-bugb in #44
• fix(hypothesis): bound every next renderer by -n, name the queue total, and key each entry by @Shahid-BugB in #43
• Prepare 2.1.0: the parser now reads the annotations 2.0.0 silently dropped by @Animesh-Sri-bugb in #45

Full Changelog: v2.0.0...v2.1.0